microsoft/onnxruntime

Pull Requests Commits

Address code review feedback: improve comments and remove model-specific references

Copilot committed 3 days ago

176245eb

Add LinearAttention contrib operator with CPU and CUDA implementation

Copilot committed 3 days ago

50b72884

Only validate REACT_NATIVE if required (#27827)

eserscor committed 4 days ago

Verified 44ce4f5f

webgpu: Fix multi-session GPU profiling event attribution (#27689)

qjia7 committed 4 days ago

Verified fb19cebe

Fill CUDA EP opset gap for Shape operator (23→25) (#27734)

Copilot committed 4 days ago

Verified 200e9f09

Update CUDA TopK kernel registration to opset 24 with BFloat16 support (#27735)

Copilot committed 4 days ago

Verified a8424e67

[CPU] Handle ONNX domain Gelu and HardSigmoid activations in the NCHWc transformer suite (#27821)

hariharans29 committed 4 days ago

Verified 1b982ddc

Fix WebGPU device destroyed on session release, breaking session recreation (#27634)

nico-martin committed 4 days ago

Verified 2f668789

[WebGPU] Einsum fixes for 5D tensors (#27779)

xenova committed 4 days ago

Verified 142eccab

Make WebGPU EP compatible with EP API (#26907)

fs-eire committed 4 days ago

Verified 99c5dd88

[MLAS] Add fused Silu and Gelu kernels for AVX512 (#27690)

hariharans29 committed 4 days ago

Verified 38a2625e

Bump flatted from 3.3.3 to 3.4.2 in /js (#27799)

dependabot[bot] committed 4 days ago

Verified faad20f9

Bump flatted from 3.3.3 to 3.4.2 in /js/react_native/e2e (#27785)

dependabot[bot] committed 4 days ago

Verified 793153bc

Bump rollup from 4.35.0 to 4.59.0 in /js/web/test/e2e/exports/testcases/vite-default (#27463)

dependabot[bot] committed 4 days ago

Verified 883b461c

Extend DQ→MatMulNBits fusion to support Gemm + per-tensor/per-channel quantization (#27769)

jambayk committed 4 days ago

Verified 37b863c6

fix webnn test case for webgpu ep: 'transpose float32 1D constant tensor default op… (#27773)

guschmue committed 5 days ago

Verified 16b556d3

fix webnn/where complicance tests for webgpu (#27776)

guschmue committed 5 days ago

Verified 0c3e5fcf

Fix NeonFp16DequantB8Bit reference to match kernel fp16 precision (#27812)

jambayk committed 5 days ago

Verified c1719194

Fix QNN SDK version propagation in Linux ort-qnn wheel build (#27800)

derdeljan-msft committed 5 days ago

Verified 36c962fd

[CUDA] Extend Pad support through opset 25 with wrap mode (#27774)

Copilot committed 5 days ago

Verified fde4e032

[CUDA] RoiAlign for opset versions 16 and 22 (#27646)

tianleiwu committed 5 days ago

Verified 45b5900d

Fix non-ASCII Unicode model path crash across session and provider code (#27724)

sagarbhure-msft committed 5 days ago

Verified 3bb032e8

[CPU/CUDA EP] Add DeformConv op support (#27393)

ShirasawaSama committed 6 days ago

Verified 163f6149

Fix tf.function retracing in TensorFlow benchmark (#27665)

Rishi-Dave committed 6 days ago

Verified 727256f8

[CPU] Update the NCHWc transformer to handle more patterns (#27691)

hariharans29 committed 6 days ago

Verified aa6f2e33

[Core] Add SpaceToDepth fusion pattern (#27747)

hariharans29 committed 6 days ago

Verified 3fa42051

Add opset 21/23 CUDA kernel registrations for Flatten, Identity, If, Loop, Scan, ConstantOfShape, Size (#27728)

Rishi-Dave committed 7 days ago

Verified ce3158ac

Bump flatted from 3.2.7 to 3.4.2 in /onnxruntime/test/wasm (#27787)

dependabot[bot] committed 8 days ago

Verified b1c5e044

Fix out of bounds read issue in cross.entropy.cc (#27568)

vraspar committed 8 days ago

Verified 0aa6adb5

Fill CUDA EP opset gap for GRU operator (14 → 22) (#27738)

Copilot committed 8 days ago

Verified 830a29ee

Older