Commits microsoft/onnxruntime

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

Copilot committed 134 days ago

Verified a4f40c11

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

daijh committed 134 days ago

Verified 1053327e

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

Rishi-Dave committed 134 days ago

Verified 359d9ab4

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

tianleiwu committed 134 days ago

Verified d2836a8f

Fix CUDA build with contrib ops disabled (#28554)

Copilot committed 134 days ago

Verified 5003d932

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

tianleiwu committed 134 days ago

Verified 43989a71

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

theHamsta committed 134 days ago

Verified b850fcbb

Parallelize CPU ScatterElements kernel via ThreadPool (#28588)

Copilot committed 134 days ago

Verified e3655b35

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

apsonawane committed 135 days ago

Verified 1e80c291

Add component governance manifest for WebGPU EP (#28599)

adrastogi committed 135 days ago

Verified e10b9a81

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

maxwbuckley committed 135 days ago

Verified cf8e4f55

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

apsonawane committed 135 days ago

Verified ee444bdc

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

adrianlizarraga committed 135 days ago

Verified 158bdef0

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

edgchen1 committed 135 days ago

Verified 84d03c58

Address MaxUnpool shortcomings msrc116345 (#28550)

yuslepukhin committed 135 days ago

Verified a894893f

Move `OrtEp` sanity checks to plugin EP creation and remove check for `OrtEp::ort_version_supported` upper bound (#28601)

edgchen1 committed 135 days ago

Verified 67156b8a

Fix Whisper genai_config context_length using wrong config attribute (#28600)

jiafatom committed 135 days ago

Verified 86aae8c0

QMoE CUDA: Rename build options, refactor PrePack, add GPU kernels (#28583)

tianleiwu committed 135 days ago

Verified ae88f9bc

CPU GroupQueryAttention: Quantized KV Cache with SIMD-optimized MLAS kernels (#28578)

tianleiwu committed 135 days ago

Verified 56640661

Fail loudly when MatMulNBits receives unsupported block_size on CPU EP (#28590)

Copilot committed 136 days ago

Verified 4236d6e9

feat(orttraining): add CPU fallback for FusedAdam optimizer (#28233)

Rishi-Dave committed 136 days ago

Verified 5996a1ec

fix: propagate output_dtype attribute when inserting Q after DQ (#28144)

Rishi-Dave committed 136 days ago

Verified 559162f2

fix: guard ORT_USE_CPUINFO on __linux__ to fix FreeBSD build (#28362)

Rishi-Dave committed 136 days ago

Verified 8d644e5d

Fix EpGraphAssignmentInfo missing issue on nodes assigned to CPU-EP by InsertCastTransformer (#28268)

qti-hungjuiw committed 136 days ago

Verified 856c546c

[CoreML EP] Add Identity, Ceil, Tile builders + drop trivial-only partitions (#28293)

maxwbuckley committed 136 days ago

Verified d0203e1c

Add linux-arm64 binaries to the foundry nuget package (#28579)

baijumeswani committed 136 days ago

Verified bdcd0298

Derive ORT_TEST_VERBOSE from System.Debug via runtime coalesce in plugin test pipelines (#28517)

edgchen1 committed 136 days ago

Verified 13630176

Add quantized KV cache support in CPU GroupQueryAttention (#28576)

tianleiwu committed 136 days ago

Verified 6fbda6d8

add more ORT_ENFORCE when building the tree structure of TreeEnsemble operators (#27677)

xadupre committed 136 days ago

Verified f6e8b0b8

QMoE CUDA EP — FP4/FP8/WFP4AFP8 Quantized Mixture-of-Experts + MoE GEMM Refactor (#28467)

tianleiwu committed 136 days ago

Verified 548ab6e5