microsoft/onnxruntime

Pull Requests Commits

tianleiwu committed 22 days ago

e9be0aad

Handle rank-one scale selection edge case

Copilot committed 29 days ago

Verified e8edaa65

Fix SimplifiedLayerNorm scale shape validation

Copilot committed 29 days ago

Verified 8cff11c2

Copilot committed 29 days ago

Verified f4c6f278

[WebGPU] Optimize MatMulNBits wide-tile with subgroup shuffle (#31703)

daijh committed 29 days ago

Verified 1bc68c1d

[WebGPU] Add robustness provider option (#31971)

qjia7 committed 29 days ago

Verified 4576ea19

[WebNN EP] Add output rank validation (#31708)

Honry committed 29 days ago

Verified 8bbb81ee

[WebGPU plugin EP] Bump plugin-ep-webgpu/VERSION_NUMBER to 0.4.0. (#31960)

edgchen1 committed 29 days ago

Verified c675b44c

[MLAS] Add an AVX-512 fused kernel for LinearAttention (#31674)

mirounga committed 29 days ago

Verified a188f435

Link delayimp.lib for onnxruntime_pybind11_state when delay-loading is enabled (#31637)

jnagi-intel committed 29 days ago

Verified 71ff5816

Update cuda plugin pipelines (#32021)

tianleiwu committed 29 days ago

Verified 21d35ea9

Filter display adapters that are using the Microsoft Basic Render Driver (#32006)

skottmckay committed 29 days ago

Verified 74c8ca5b

Handle zero-element input/bias in BiasGelu and FastGelu as a no-op (#31698)

titaiwangms committed 30 days ago

Verified dfb403d9

[Test] Enable CUDA coverage for shared PagedAttention contrib-op tests via cache aliasing harness (#31687)

hariharans29 committed 30 days ago

Verified 10008a65

Fix stale ContribOperators.md for MRotaryEmbedding (#31985)

justinchuby committed 30 days ago

Verified f2dfa4e9

[WebGPU] Add Gather int64 support and make kernel version numbers function params (#31714)

miaobin committed 30 days ago

Verified faf29fef

Validate input rank/size for BifurcationDetector before indexing (#31701)

titaiwangms committed 30 days ago

Verified bbf20701

Guard three graph-optimizer passes against unbounded model-supplied indices (#31670)

titaiwangms committed 30 days ago

Verified f25933d5

Reject STRING data_type in LoRA adapter parameter loading (#31682)

apsonawane committed 30 days ago

Verified f5bfd940

Bump ORT Version to 1.30 (#31972)

tianleiwu committed 30 days ago

Verified 5dd1462b

[CUDA] Add GatedAdd contrib operator (#31835)

tianleiwu committed 30 days ago

Verified 9b174a42

Validate SkipLayerNorm prepacked lengths (#31676)

apsonawane committed 30 days ago

Verified a29da166

Validate whisper beginning_timestamp_token_id bounds (#31636)

apsonawane committed 30 days ago

Verified 831fa2fe

Handle RNN activation parameters safely (#31675)

apsonawane committed 30 days ago

Verified 3cee1860

Harden CUDA fp16 transpose index math against int overflow (#31644)

apsonawane committed 30 days ago

Verified 37d79e38

Validate QLinearConv bias size against output channels (#31684)

apsonawane committed 30 days ago

Verified 45bdd4c4

Fix TfIdfVectorizer weight indexing semantics (#31649)

apsonawane committed 30 days ago

Verified f5de4c39

Validate FeatureVectorizer batch sizes (#31671)

apsonawane committed 30 days ago

Verified d3abafc3

[WebGPU] Add MRotaryEmbedding support (#31976)

qjia7 committed 30 days ago

Verified 1441b36d

Fix LinearAttention on GPUs with limited shared memory (#31982)

Copilot committed 30 days ago

Verified 86d38813

Older