Commits microsoft/onnxruntime

Address CUDA ReduceL2 negative results and add test coverage (#28681)

yuslepukhin committed 69 days ago

Verified efbffdd2

[WebGPU] QKV and MLP layer fusions for Qwen3-style models (#28280)

hariharans29 committed 69 days ago

Verified 88ca23fc

Update protobuf version to 6.33.5 (#28684)

sushraja-msft committed 69 days ago

Verified 04c900d6

[webgpu] Per-graph buffer manager for WebGPU multi-graph capture (#28260)

qjia7 committed 70 days ago

Verified 5d6cbb67

Update GQA op schema (#28689)

apsonawane committed 70 days ago

Verified 46fdcbba

Use fp32 accumulation in SkipLayerNorm/EmbedLayerNorm CUDA kernels (#28682)

tianleiwu committed 70 days ago

Verified 26871ab4

[WebGPU plugin EP] Enforce minimum ORT version specified by `plugin-ep-webgpu/MIN_ONNXRUNTIME_VERSION` (#28687)

edgchen1 committed 70 days ago

Verified b1880617

Guard removed APIs in TRT 11 (#28611)

kevinch-nv committed 70 days ago

Verified 53fcead6

[CoreML EP] Add Sin and Cos unary ops (#28596)

maxwbuckley committed 70 days ago

Verified 71cfbb0f

Speed up CUDA CI build: split into per-arch OBJECT libraries, add --flash_nvcc_threads, and enable quick build mode (#28645)

tianleiwu committed 71 days ago

Verified b0924f47

Propagate outer scope type info to subgraphs during verification (#27707)

Rishi-Dave committed 71 days ago

Verified bdd52fa1

Fix CUTLASS FMHA BiasLoader alignment for unaligned kernel path (#28369)

justinchuby committed 71 days ago

Verified 1f92c9d8

remove some webgpu ci tasks to speedup ci (#28642)

Guenther Schmuelling committed 71 days ago

Verified a42351e3

Check weight shape dimensions in ConvTranspose shape inference msrc116345 (#28524)

yuslepukhin committed 71 days ago

Verified 12916e9e

Add --test_parallel to reduce CTest concurrency for ASan builds (#28675)

tianleiwu committed 71 days ago

Verified c85f6eb9

Fix plugin EP test failure when host ORT lacks newer data types (#28659)

tianleiwu committed 71 days ago

Verified 2c4a0c18

lstm support for webgpu (#27881)

Guenther Schmuelling committed 71 days ago

Verified 30b6528f

rev some npm packages (#28547)

Guenther Schmuelling committed 71 days ago

Verified e0a8dac2

Update GatherBlockQuantized to support 2-bits (#28530)

HectorSVC committed 72 days ago

Verified 97cfc86f

Raise protobuf minimum version in Python requirements (#28558)

anzzraju1997-glitch committed 72 days ago

Verified 7a3f9a21

[Annotations] Address cuda orginal matching (#28623)

yuslepukhin committed 72 days ago

Verified b5fb2b89

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

Copilot committed 72 days ago

Verified c267f8e2

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

tairenpiao committed 72 days ago

Verified c5afcc5d

Additional diagnostics for DML failure path (#28495)

adrastogi committed 74 days ago

Verified a1fc9165

Add example and documentation for kOrtEpDevice_EpMetadataKey_OSDriverVersion (#28282)

adrastogi committed 75 days ago

Verified d464b2a0

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

velonica0 committed 75 days ago

Verified b2f6e151

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

jiafatom committed 75 days ago

Verified a4f79e82

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

Copilot committed 75 days ago

Verified a4f40c11

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

daijh committed 75 days ago

Verified 1053327e

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

Rishi-Dave committed 75 days ago

Verified 359d9ab4