microsoft/onnxruntime

Pull Requests Commits

wip: add support for using pre-packed weights with CPU matmulnbits impl

derdeljan-msft committed 27 days ago

be075e4f

Fix issue w/ DML batched readback size (#28680)

adrastogi committed 29 days ago

Verified 90c095d1

fix: require GCC >= 12 on aarch64 for bf16/fp16 capability probes (#28785)

Rishi-Dave committed 29 days ago

Verified d9ecd81d

Fix INT64 initializer data propagation (#28778)

GopalakrishnanN committed 29 days ago

Verified 0b7823e0

Validate Col2Im inputs to prevent heap over-read (#28706)

GopalakrishnanN committed 29 days ago

Verified 6f589d2a

Fix thread-pool param self-assignment and dedupe schema registration in session_state_test (#28799)

GopalakrishnanN committed 29 days ago

Verified dfaf4b44

creates zips to consume cuda plugin ep from foundry local (#28789)

prathikr committed 29 days ago

Verified e6c16dff

Fix missing cuDNN DLL preload from NVIDIA site packages on Windows (#28787)

badranX committed 30 days ago

Verified 47439360

Fix CI: update import convert_tf_models_to_pytorch (#28798)

tianleiwu committed 30 days ago

Verified 69018c1d

Fix NHWC second-pass EP assignment cleanup (#27858)

tianleiwu committed 30 days ago

Verified 646540ec

Address resize shortcomings (#28779)

yuslepukhin committed 30 days ago

Verified 84eb4f6e

Clarify Kahan condition (#28790)

yuslepukhin committed 30 days ago

Verified 4eb757c7

Fix ASan OOM in QDQ Gemm transformer tests (#28797)

tianleiwu committed 30 days ago

Verified d43a48bd

Python API for HOST_ACCESSIBLE OrtValue allocation (#28038)

ericcraw committed 31 days ago

Verified b3e1a9e6

Switch NPM publishing to consume from CUDA 13 pipeline (#28773)

kpkbandi committed 31 days ago

Verified d949c346

Model Package Support (Phase 2) (#28271)

chilo-ms committed 31 days ago

Verified 27c6d31b

fix(transformers): validate archive extraction paths (#28777)

tianleiwu committed 31 days ago

Verified f1f04362

fix(ci): use subprocess for Triton compile helpers (#28775)

tianleiwu committed 31 days ago

Verified 368634ee

fix: avoid shell injection in training helper (#28776)

tianleiwu committed 31 days ago

Verified 7f326735

Harden CropAndResize against malformed crop_size tensors (#28766)

tianleiwu committed 31 days ago

Verified 2603ef50

Validate BeamSearch vocab_size against logits width (#28774)

tianleiwu committed 31 days ago

Verified 211169cc

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

tairenpiao committed 31 days ago

Verified 1686e088

Address PadFusion shortcomings (#28780)

yuslepukhin committed 31 days ago

Verified eded2c74

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

Rishi-Dave committed 31 days ago

Verified b40bdd80

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

Rishi-Dave committed 31 days ago

Verified e6b3eb0e

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

Copilot committed 31 days ago

Verified 03245616

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

Copilot committed 31 days ago

Verified 55975681

Fix regression in RandomForestClassifier binary classification predictions (#28685)

Copilot committed 31 days ago

Verified db641d65

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

Rishi-Dave committed 31 days ago

Verified e2789cdf

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

skottmckay committed 31 days ago

Verified 72138269

Older