Commits microsoft/onnxruntime

Merge remote-tracking branch 'origin/main' into copilot/update-transpose-cuda-operator

tianleiwu committed 7 days ago

58955c73

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

guschmue committed 7 days ago

Verified 0c3e5fcf

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

jambayk committed 7 days ago

Verified c1719194

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

derdeljan-msft committed 7 days ago

Verified 36c962fd

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

Copilot committed 7 days ago

Verified fde4e032

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

tianleiwu committed 7 days ago

Verified 45b5900d

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

sagarbhure-msft committed 7 days ago

Verified 3bb032e8

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

ShirasawaSama committed 8 days ago

Verified 163f6149

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

Rishi-Dave committed 8 days ago

Verified 727256f8

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

hariharans29 committed 8 days ago

Verified aa6f2e33

[Core] Add SpaceToDepth fusion pattern (#27747)

hariharans29 committed 8 days ago

Verified 3fa42051

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

Rishi-Dave committed 9 days ago

Verified ce3158ac

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

dependabot[bot] committed 10 days ago

Verified b1c5e044

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

vraspar committed 10 days ago

Verified 0aa6adb5

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

Copilot committed 10 days ago

Verified 830a29ee

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

kevinlam92 committed 10 days ago

Verified b10fa545

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

eserscor committed 10 days ago

Verified e3ebf875

[WebGPU] DequantizeLinear op fixes (#27706)

edgchen1 committed 10 days ago

Verified a6857897

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

Copilot committed 11 days ago

Verified f546db86

Add SkipLayerNorm fusion with bias Add (#27765)

kunal-vaishnavi committed 11 days ago

Verified 405dcd72

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

tianleiwu committed 11 days ago

Verified 988d05e4

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

jambayk committed 11 days ago

Verified d9bb7602

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

titaiwangms committed 11 days ago

Verified 93d31cfd

Java: avoid provider resource extraction when library already exists in onnxruntime.native.path (#27668)

TsofnatMaman committed 11 days ago

Verified c0fd2f78

Fix integer division by zero crash in CPU EP Div operator (#27693)

Copilot committed 11 days ago

Verified 44e8b384

Fill Squeeze and Unsqueeze CUDA opset gaps to opset 25 (#27739)

Copilot committed 11 days ago

Verified 72d46059

Improve checks in concat_slice_elimination against optional attrs and unsqueeze_elimination against invalid model (#27638)

yuslepukhin committed 11 days ago

Verified d0b7cf2e

Add optional router_weights input to QMoE for separate selection/aggregation routing (#27687)

Copilot committed 12 days ago

Verified 5bba3573

Fix onnxruntime-qnn python wheel build on WSL (#27730)

derdeljan-msft committed 12 days ago

Verified c896904f

Add ORT_ENFORCE for swiglu_fusion==1 in QMoECPU to prevent out-of-bounds read (#27748)

Copilot committed 12 days ago

Verified ba654836