microsoft/onnxruntime

Pull Requests Commits

Copilot committed 14 hours ago

Verified e585b7ae

[CUDA] Add FP16-cache paged XQA decode for head size 256 (#32263)

baijumeswani committed 1 day ago

Verified b1f76d58

Fix ort-ci/SKILL.md yaml frontmatter syntax. (#32248)

edgchen1 committed 1 day ago

Verified 05491530

Harden head-size-256 paged XQA coverage (#32262)

baijumeswani committed 2 days ago

Verified d67cc269

Run wgsl_template Python tests in CI (#32214)

edgchen1 committed 2 days ago

Verified dc999734

[MLAS] Reject Arm® KleidiAI™ Q4 prepack with dynamic scales (#32068)

martin-klacer-arm committed 3 days ago

Verified 9cbf665b

[MLAS] Refactor QNBit KleidiAI integration into dedicated layer (#31720)

martin-klacer-arm committed 3 days ago

Verified 5cb241b5

Validate MatMulNBits 8-bit g_idx bounds on CUDA (#31643)

apsonawane committed 3 days ago

Verified 066b90ef

Fix Windows CUDA 12.9 SM120 compilation (#32114)

Copilot committed 3 days ago

Verified 939e180a

pin review dog (#32208)

eserscor committed 3 days ago

Verified ddc8f0c3

Upgrade Protobuf to v33.6 (#29906)

Copilot committed 3 days ago

Verified 993cd135

Extend memory importer with host pointer support (#29726)

javier-intel committed 3 days ago

Verified 710f0ce3

[Telemetry] Prevent crashes in shell-less containers (#32226)

bmehta001 committed 3 days ago

Verified 9aa8fb06

[CUDA] Add VarlenCausalConvWithState for continuous batching (#32168)

baijumeswani committed 3 days ago

Verified 2a9bf1e9

Support LoRA adapters with plugin EP allocators (#32221)

baijumeswani committed 3 days ago

Verified 4f7d0f5b

[CUDA] Extend paged XQA decode to head size 256 (#32229)

baijumeswani committed 3 days ago

Verified 15920d08

Move ort-release-notes skill to new .github/skills location. (#32193)

edgchen1 committed 3 days ago

Verified ddbcef7d

[WebGPU] Supply fused activation parameters to Conv/MatMul as uniforms (#32116)

4n4ny4 committed 5 days ago

Verified 4d308dac

[WebGPU] Scale Dawn pipeline compilation workers with CPU count (#29820)

xiaofeihan1 committed 6 days ago

Verified 89b01b40

Fix C API docs Doxygen download (#32210)

Wayne-Ch committed 6 days ago

Verified df012c2f

[ARM] MLAS: SVE i8mm (svmmla) int8 QGEMM kernels, portable machine code (#31146)

mirounga committed 6 days ago

Verified a4156be6

[MLAS] Add a NEON fused kernel for LinearAttention (#32178)

mirounga committed 6 days ago

Verified 27f212c4

[WebGPU] Regenerate WGSL in-tree golden fixtures to match current templates (#32115)

4n4ny4 committed 6 days ago

Verified 775526b8

[WebGPU] Make MatMulNaiveProgram's pipeline cache key cover everything it bakes into WGSL (#32048)

4n4ny4 committed 6 days ago

Verified 8772f770

Pin GitHub Actions to full-length commit SHAs (#32176)

danfiedler-msft committed 6 days ago

Verified 5fd8e199

Resolve possible onnx CVE in python docs (#32190)

eserscor committed 6 days ago

Verified 1fbac8f3

[MLAS] AVX-512 16-wide Erf kernel and NCHWc reorder transpose for MobileClip-S0 model (#31958)

swetha097 committed 7 days ago

Verified a88c1168

[WebGPU] Optimized PagedAttention implementation (2/n) (#31727)

hariharans29 committed 7 days ago

Verified ebd7bad1

[MLAS] Hardswish Fusion implement for Mobilenetv3 models (#31957)

swetha097 committed 7 days ago

Verified 4db33198

Rework agent code review guidance (#31969)

edgchen1 committed 7 days ago

Verified 5d9676ea

Older