microsoft/onnxruntime

Pull Requests Commits

Add staged CI pipeline: Stage 2 (Linux CUDA) waits for Linux CI, Stage 3 (all others) wait for Linux CUDA CI

Copilot committed 101 days ago

Verified 2c7201aa

Copilot committed 101 days ago

Verified 192e80de

Bound total output allocation size in Tile kernel (#28070)

GopalakrishnanN committed 101 days ago

Verified a363e895

Fix OOB reads in SoftmaxCrossEntropyLoss via label bounds validation (#28004)

vraspar committed 101 days ago

Verified ad11f7cf

Use weights_only=True for remaining torch.load() calls (#28421)

adrianlizarraga committed 101 days ago

Verified 4e377a13

Add LabelEncoder_4 fast path for session creation (#27526)

SamuelLess committed 101 days ago

Verified 73d54165

Add RVV (RISC-V Vector Extension) optimized convolution and pooling kernels for the NCHWc blocked format in MLAS (#28411)

velonica0 committed 101 days ago

Verified 0e721889

Address Resize kernel shortcomings (#28402)

yuslepukhin committed 101 days ago

Verified 469bc9f6

[webgpu] Optimize LinearAttention Op with subgroup (#28412)

daijh committed 101 days ago

Verified a8260cd6

Fill CUDA EP opset gap for Sin and Cos operators (7→22) (#27756)

Copilot committed 101 days ago

Verified 480aad09

Use symbol file to export symbols in onnxruntime library on AIX platform as well (#28016)

ayappanec committed 101 days ago

Verified 239ef9e3

[WebGPU] Correct MatMul bias input indexing in `MatMul::ComputeInternal` (#28475)

Copilot committed 101 days ago

Verified 72bce1df

fix: skip DQ->MatMulNBits fusion when weight/scale initializer is shared (#28326)

Rishi-Dave committed 101 days ago

Verified 183e7a94

[WebGPU] Fix SkipSimplifiedLayerNormalization bias (#28427)

xenova committed 101 days ago

Verified fafe5644

Add CPU QMoE 2-bit support and LUT GEMM fast path (#28185)

tianleiwu committed 101 days ago

Verified dac5a143

[ Cuda] ConvTranspose-22 (#27710)

tianleiwu committed 102 days ago

Verified a12b611e

Fix sigmoid transformation in TreeEnsembleClassifier for all-positive weights with LOGISTIC post_transform (#27536)

Copilot committed 102 days ago

Verified 756a5235

Enable QuickGeluFusion on WebGPU EP and fix fp16 shader (#28410)

xiaofeihan1 committed 102 days ago

Verified a49da870

Fix runtime-unresolvable type annotations in Session and InferenceSession (#27802)

Rishi-Dave committed 102 days ago

Verified 8d24011e

Add DiT attention fusion for F5-TTS and diffusion transformer models (#27999)

Rishi-Dave committed 102 days ago

Verified ece097dc

Add CPU Cast op support for Float8E8M0 (#28435)

tianleiwu committed 102 days ago

Verified 8cfb63e0

[CUDA] PagedAttention: use exact max_query_len on FA path (#28409)

elwhyjay committed 102 days ago

Verified d82a06df

[GQA] Make present_key/present_value outputs optional and add Gemma4 support (#28242)

apsonawane committed 102 days ago

Verified 83e402e0

Validate NodeArg creation (#28429)

yuslepukhin committed 102 days ago

Verified 51c546b0

Fix CUDA plugin EP test pipeline (#28453)

tianleiwu committed 102 days ago

Verified c1a61b44

Address tokenizer shortcomings (#28428)

yuslepukhin committed 102 days ago

Verified 67d9400d

Fix Clang + libc++ compilation and portability issues (#28049)

mustjab committed 102 days ago

Verified c6d9a020

Handle empty initializers gracefully in optimizer passes (#27976)

GopalakrishnanN committed 102 days ago

Verified 1b64fba3

[WebGPU Plugin EP] Packaging pipelines minor updates (#28460)

edgchen1 committed 102 days ago

Verified 0cd3de86

Harden OneHot operator input validation and output size computation (#28014)

GopalakrishnanN committed 103 days ago

Verified df114575

Older