microsoft/onnxruntime

Pull Requests Commits

Fix comment formatting

adrastogi committed 2 days ago

4ea536f8

Limit buffer reuse guidance to caller-owned memory

adrastogi committed 2 days ago

92428d08

Document that initializers added with data_is_external=true are read-only

adrastogi committed 2 days ago

bb98066e

[web] Exclude graph and ONNX schema code from Asyncify (#33563)

El1as42 committed 2 days ago

Verified 1a6aae36

[perftest] Enable profiling for the compile session with --compile_ep_context (#33694)

MayureshV1 committed 2 days ago

Verified ee7f7c36

Stage Dawn Agility SDK through configured NuGet feed (#33648)

edgchen1 committed 2 days ago

Verified c4c0e001

Validate types before MatMulIntegerToFloat fusion (#33485)

xgbtxy committed 2 days ago

Verified 0a12bc04

Handle empty batch-transposed MatMul operands (#33512)

rookieLiu2018 committed 2 days ago

Verified 398e46bf

Order NaN as the largest value in the CPU TopK kernel (#33543)

HuzaifaChaudary committed 2 days ago

Verified d80d43cd

Fix unused KleidiAI kernel name in builds without diagnostic consumers (#33583)

leninworld committed 2 days ago

Verified b32ab10d

Fix fp16 grouped Conv packed-weight offsets and per-group bias on arm64 (#33598)

usmanmateen committed 2 days ago

Verified af2449eb

Fix QDQ MatMul fusion by validating quantization parameters (#33616)

the0cp committed 2 days ago

Verified f8304d45

MatMulNBits subgroup matrix 8x16x16: reduce robustness overhead (#33674)

jchen10 committed 2 days ago

Verified 0dacf7ee

Fix CUDA training stream API callers (#32981)

Yuhx141 committed 2 days ago

Verified 123752d3

Centralize Component Governance exclusions across pipelines (#33612)

eserscor committed 2 days ago

Verified 9f807084

Enforce GitHub Actions SHA pinning in lint checks (#33658)

eserscor committed 2 days ago

Verified e2019ee8

Bump actions/download-artifact from 8.0.1 to 8.0.2 (#33387)

dependabot[bot] committed 3 days ago

Verified 526e95a1

[WebGPU] Return empty pooling outputs without dispatching (#33668)

gyagp committed 3 days ago

Verified 9fd928c2

[WebGPU] Fix Gemm with rank-0 scalar bias (#33664)

gyagp committed 3 days ago

Verified 85c8b6dd

[CUDA] Optimize INT3 QMoE decode/prefill kernels and S=8 GEMV finalize dispatch (#33585)

tianleiwu committed 3 days ago

Verified c026a509

[CUDA] Extend M=8 narrow CtaN=2 in fpA_intB_gemv to BF16 and INT8 small-N projections (#33600)

tianleiwu committed 3 days ago

Verified f30f7fcb

[WebGPU] Use Queue.WriteBuffer for small CPU-to-GPU uploads (#33551)

aciddelgado committed 3 days ago

Verified 49481a71

[WebGPU] Batch streamless GPU-to-GPU copy vectors (#33567)

aciddelgado committed 3 days ago

Verified af0a4bab

[CUDA] Optimize GatedDeltaNetChunkedKernel with ldmatrix/SyncBeforeStore and add GatedRMSNormWarp128Kernel (#33601)

tianleiwu committed 3 days ago

Verified 1e3c937d

Update Pillow minimum version to 13.0.0 (#33656)

mustjab committed 3 days ago

Verified 7201c965

Deprecate provider-bridge legacy NvTensorRTRTXExecutionProvider (#32276)

keshavv27 committed 3 days ago

Verified 979b6d1d

[WebGPU] Optimize DP4A for large Qualcomm subgroups (#33560)

gyagp committed 3 days ago

Verified 034e0eaa

[CUDA] Include SM103 in Linux aarch64 CUDA 13 packages (#33649)

haoxiz-nvidia committed 3 days ago

Verified edb399aa

[WebGPU] Fix subgroup matrix configs not populated in WASM builds (#33531)

kylo5aby committed 3 days ago

Verified c130fb69

[CPU/WebGPU] Support empty MatMulNBits weights (#33559)

qjia7 committed 3 days ago

Verified 03a19a74

Older