microsoft/onnxruntime

Pull Requests Commits

Fix FURB110 lint: use `or` instead of ternary for v_head_size

titaiwangms committed 44 days ago

0bdde29d

Add asymmetric head_size regression test for MEA fallback

titaiwangms committed 44 days ago

aadf5da1

Zero present buffers before concat to prevent NaN propagation

titaiwangms committed 44 days ago

c2da4b12

Fix test review findings for MEA+decode tests

titaiwangms committed 44 days ago

7318e243

Fix review findings: use v_head_size for V ops, add safety comment

titaiwangms committed 44 days ago

53bddd27

Fix MEA eligibility: skip decode when head_size != v_head_size

titaiwangms committed 44 days ago

f5befa80

Add MEA+decode support in ONNX Attention LLM op

titaiwangms committed 44 days ago

6ab008b7

Add MEA+decode test cases for ONNX Attention LLM op

titaiwangms committed 44 days ago

719964ae

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

guschmue committed 46 days ago

Verified 0c3e5fcf

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

jambayk committed 46 days ago

Verified c1719194

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

derdeljan-msft committed 46 days ago

Verified 36c962fd

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

Copilot committed 46 days ago

Verified fde4e032

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

tianleiwu committed 46 days ago

Verified 45b5900d

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

sagarbhure-msft committed 46 days ago

Verified 3bb032e8

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

ShirasawaSama committed 46 days ago

Verified 163f6149

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

Rishi-Dave committed 47 days ago

Verified 727256f8

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

hariharans29 committed 47 days ago

Verified aa6f2e33

[Core] Add SpaceToDepth fusion pattern (#27747)

hariharans29 committed 47 days ago

Verified 3fa42051

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

Rishi-Dave committed 48 days ago

Verified ce3158ac

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

dependabot[bot] committed 49 days ago

Verified b1c5e044

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

vraspar committed 49 days ago

Verified 0aa6adb5

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

Copilot committed 49 days ago

Verified 830a29ee

[React Native] Fix float16 tensor crash by mapping to Uint16Array (#27549)

kevinlam92 committed 49 days ago

Verified b10fa545

Dev/erscor/2026/3/19 make react native optional (#27772)

eserscor committed 49 days ago

Verified e3ebf875

[WebGPU] DequantizeLinear op fixes (#27706)

edgchen1 committed 49 days ago

Verified a6857897

Fill GlobalAveragePool and GlobalMaxPool opset gap in CUDA provider (1→22) (#27733)

Copilot committed 49 days ago

Verified f546db86

Add SkipLayerNorm fusion with bias Add (#27765)

kunal-vaishnavi committed 49 days ago

Verified 405dcd72

[CUDA] Refactor CUDA helpers or ops for plugin EP (#27719)

tianleiwu committed 49 days ago

Verified 988d05e4

Enable KleidiAI for asymmetric 4-bit MatMulNBits on ARM64 (#27751)

jambayk committed 49 days ago

Verified d9bb7602

Support softcap and softmax_precision in Attention(CUDA) (#27714)

titaiwangms committed 50 days ago

Verified 93d31cfd

Older