microsoft/onnxruntime

Pull Requests Commits

Revert "Draft: retain plugin EP compile-time initializer ownership"

chilo-ms committed 20 days ago

e993e720

Draft: retain plugin EP compile-time initializer ownership

chilo-ms committed 20 days ago

d15def57

Clarify partitioned CUDA output and failure contracts

chilo-ms committed 33 days ago

4f2f838d

Merge main and preserve partitioned CUDA failure guard

chilo-ms committed 34 days ago

e72a3171

Reject eager runs after partitioned CUDA graph failure

chilo-ms committed 34 days ago

6cd0ffd1

WebGPU static plugin EP (#32395)

edgchen1 committed 34 days ago

Verified a84958ac

Bump github/codeql-action/autobuild from 3.37.7 to 4.38.2 (#33024)

dependabot[bot] committed 34 days ago

Verified d9b8ddc0

Filter model package output options (#33096)

apsonawane committed 34 days ago

Verified b0609a04

Validate GatherBlockQuantized output dimensions (#33097)

apsonawane committed 34 days ago

Verified 003599e1

Validate DLPack capsules before conversion (#33100)

apsonawane committed 34 days ago

Verified 99f13d09

Bound antialias resize coefficient windows (#33095)

apsonawane committed 34 days ago

Verified 27800d7f

Limit AOT function expansion (#32662)

apsonawane committed 34 days ago

Verified 2f3c9186

[CUDA] Validate GEMM autotuning candidates using graph replay (#33085)

tianleiwu committed 34 days ago

Verified ba9d0320

Extend CUDA QMoE packed INT2 prefill block sizes and weight combinations (#33092)

jiafatom committed 34 days ago

Verified 0c67e77b

[CUDA] Add tinygemm2 as a GEMM auto-tune candidate for small-M MatMul on SM 9.0+ (#32901)

tianleiwu committed 34 days ago

Verified d2d1e8fe

[CUDA] Handle missing optional input tensors in CUDA kernels (#32971)

tianleiwu committed 35 days ago

Verified 5a7b7903

Add external Dawn API package CI coverage for the WebGPU plugin (#33077)

mustjab committed 36 days ago

Verified 690e7312

[WebGPU] Implement SparseAttentionIndexer operator (#32528)

Copilot committed 36 days ago

Verified befed9ad

[WebGPU] Optimize WebGPU SparsePagedAttention softmax execution (#32613)

Copilot committed 36 days ago

Verified 9edd1796

[CUDA] Optimize DynamicSparseAttention for large-context inference (#32671)

Copilot committed 36 days ago

Verified 8478fe99

Revert ONNX 1.23 integration pending upstream release fixes (#33073)

gramalingam committed 37 days ago

Verified e2363788

[WebGPU] Separate command recording from buffer reuse tracking (#32874)

qjia7 committed 37 days ago

Verified 4ae10e3e

[WebGPU] Add DynamicSparseAttention (#32529)

Copilot committed 37 days ago

Verified 0aecb6c1

[WebGPU] Add SparsePagedAttention (#32533)

Copilot committed 37 days ago

Verified cfbc790b

Synchronize telemetry state capture (#32748)

apsonawane committed 37 days ago

Verified 8fe6184f

Handle zero-channel ConvTranspose inputs (#33064)

apsonawane committed 37 days ago

Verified 8b9c6c77

Migrate more pipelines to single windows pool (#32678)

eserscor committed 37 days ago

Verified 1b1c1f21

Reject package-controlled optimized model paths (#33059)

apsonawane committed 37 days ago

Verified 982fcf87

Validate NCHWc binary input ranks (#33060)

apsonawane committed 37 days ago

Verified f537be06

Handle out-of-range antialiased Resize coordinates (#33056)

apsonawane committed 37 days ago

Verified ba9e4b10

Older