microsoft/onnxruntime

Pull Requests Commits

Reduce MaxPool QDQ test footprint for ASAN build stability

Copilot committed 68 days ago

Verified a2644016

Drop QDQ around MaxPool for opset 22

Copilot committed 68 days ago

Verified 14f08d46

Copilot committed 68 days ago

Verified 90250744

Fix CPU ReverseSequence returning zeros for zero-length sequences (#28759)

tairenpiao committed 68 days ago

Verified 1686e088

Address PadFusion shortcomings (#28780)

yuslepukhin committed 68 days ago

Verified eded2c74

docs(quantization): clarify CPU parameter recommendations (#28415)

Rishi-Dave committed 69 days ago

Verified b40bdd80

fix: support Float16/BFloat16/Float8 in TensorArray custom op (#28335)

Rishi-Dave committed 69 days ago

Verified e6b3eb0e

Fix incorrect output for QDQ Resize with non-nearest interpolation modes under ORT_ENABLE_ALL (#28454)

Copilot committed 69 days ago

Verified 03245616

Pre-check ConstantOfShape output size against input initializer before constant folding (#28751)

Copilot committed 69 days ago

Verified 55975681

Fix regression in RandomForestClassifier binary classification predictions (#28685)

Copilot committed 69 days ago

Verified db641d65

fix(quantization): skip MaxPool during FP8 static quantization (#28488)

Rishi-Dave committed 69 days ago

Verified e2789cdf

Check for duplicate and null attributes in Model Editor API CreateNode (#28758)

skottmckay committed 69 days ago

Verified 72138269

Register Flatten as Direct8Bit op in Python QDQ static quantizer (#28340)

Rishi-Dave committed 69 days ago

Verified 9984c702

Reject QDQ Gemm→QGemm fusion when alpha != 1 with bias (#28131)

elwhyjay committed 69 days ago

Verified 58527303

Address numerical accuracy issues. (#28743)

yuslepukhin committed 69 days ago

Verified 18e3e86d

chore: bump version to 1.28.0 (#28765)

sanaa-hamel-microsoft committed 69 days ago

Verified 3ddbdfba

[WebGPU] Default non-debug validation mode set to disabled (#28679)

hariharans29 committed 69 days ago

Verified e25750d8

[CPU] Add fix for fastmath corner case with failing tests via sbgemm_neon_kernel (#28394)

JonathanC-ARM committed 69 days ago

Verified e0d99e27

fix the case where the buffer for past_state == present_state (#28753)

Guenther Schmuelling committed 69 days ago

Verified e26ad80e

Skip SetupDi device discovery if Win32k system calls are disabled (#28535)

shiyi9801 committed 69 days ago

Verified a8baf5c4

fix(quantization): validate bias scale in QDQ Conv → QLinearConv fusion (#28229)

Rishi-Dave committed 69 days ago

Verified 5c34495a

Use abseil for readable POSIX stack traces in debug builds (#28405)

tianleiwu committed 70 days ago

Verified 8da5e910

feat(quantization): add opset-21 block_size attribute to QDQ (#28522)

Rishi-Dave committed 70 days ago

Verified 8fcb725d

fix: preserve Q/DQ around Resize when mode is linear or cubic (#28414)

Rishi-Dave committed 70 days ago

Verified 7abc315f

Fix PRelu returning NaN for infinite inputs on CPU EP (#28750)

Copilot committed 70 days ago

Verified 9f81e2df

[MLAS][KleidiAI] Apply fp16 split independent fixes (#28720)

Laan33 committed 70 days ago

Verified 1e2b6394

Zero-copy I/O for plugin EPs with HOST_ACCESSIBLE memory (#28037)

ericcraw committed 70 days ago

Verified a827f9c3

[CoreML EP] Add GatherND builder (#28598)

maxwbuckley committed 70 days ago

Verified 21e9fe39

Fix ARM CPUIDInfo bounds handling for unknown CPU vendors (#28344)

Copilot committed 70 days ago

Verified d1135499

Honor disable_synchronize_execution_providers for CUDA graph replay (#28686)

tianleiwu committed 70 days ago

Verified 312f5249

Older