microsoft/onnxruntime

Pull Requests Commits

Copilot committed 5 days ago

Verified 95684962

Reject non-control-flow nodes that carry subgraphs during session state finalization (#32641)

miaobin committed 5 days ago

Verified b95cf781

Batch QMoE CPU QNBit expert GEMMs into grouped MLAS dispatches (#32668)

tianleiwu committed 5 days ago

Verified 4b728dd1

Reject nodes that feed inputs to a zero-input operator schema (#32633)

shiyi9801 committed 5 days ago

Verified 613bc03d

Skip redundant present_key/value copy when aliased to external KV cache (follow-up to #29715) (#31150)

titaiwangms committed 5 days ago

Verified c7b64756

Add review guidance for deprecated components (#32620)

edgchen1 committed 5 days ago

Verified b01a4431

Fix WebNN ConvTranspose shape arithmetic overflow (#32661)

edgchen1 committed 5 days ago

Verified 2624ff1c

[objc docs] Fix doc generation workflow (#32550)

edgchen1 committed 5 days ago

Verified 0d453b23

Verify Level-2 workspace reservations (#32189)

chilo-ms committed 5 days ago

Verified e10b9dfc

[MLAS] Riscv64 rvv kernel optimizations (#32540)

Zestion committed 5 days ago

Verified a76f0666

[WebGPU] Enable subgroup matrix kernels on AMD GPUs (#32667)

gyagp committed 5 days ago

Verified bff4fcf7

[Plugin EP] Preserve streams in single-tensor data transfers (#32666)

xiaofeihan1 committed 5 days ago

Verified 5899c7c9

[CANN] Fix deadlock during parallel execution (#29740)

abeclowicz committed 5 days ago

Verified 74e915f9

Fix out-of-bounds read in EmbedLayerNormalization shape inference for scalar beta (#32610)

wangw-1991 committed 5 days ago

Verified 70546574

perf: run block-wise int4/int8 QMoE CPU experts on the MLAS QNBit GEMM kernels (#32644)

ykhrustalev committed 5 days ago

Verified c18808b9

Fix QuickGelu fusion for unsupported CPU dtypes (#32427)

sylvesterkaczmarek committed 6 days ago

Verified 8d399fe7

Restrict XQA CUDA compilation to SM80 and newer (#32547)

Copilot committed 6 days ago

Verified 0db6c0e8

Improve CPU provider build time with Eigen PCH (#32616)

titaiwangms committed 6 days ago

Verified f9b87e94

Add BF16 routing support to CUDA MoE (#32659)

kunal-vaishnavi committed 6 days ago

Verified e266b17e

Add complete GQA workspace route recipes (#32454)

titaiwangms committed 6 days ago

Verified 8808c862

Optimize CPU uint8 GatherBlockQuantized dequantization (#32621)

Copilot committed 6 days ago

Verified 180041af

Validate NCHWc convolution layouts (#32557)

apsonawane committed 6 days ago

Verified 460dee98

Fix --enable_arm_neon_nchwc being a silent no-op on Windows ARM64 (#32094)

Copilot committed 6 days ago

Verified f38538cd

[CUDA] Add opt-in SM90 DeepGEMM for MatMulBlockQuantizedFp8Weight (#32599)

tianleiwu committed 6 days ago

Verified 29874882

Minor follow-up to GQA value tensor layout changes (PR 32139) (#32519)

javier-intel committed 6 days ago

Verified 30e26321

[WebGPU] Move several elements per thread in Concat, Slice, Transpose and Pool (#32497)

linzuojian committed 6 days ago

Verified 2a720f92

[CUDA] Make GatherND capture-safe (#32120)

tianleiwu committed 7 days ago

Verified 71b33460

Specify GroupQueryAttention windowed KV cache layout contract (#32419)

tianleiwu committed 7 days ago

Verified cd9da083

[CUDA] Add cuDNN paged SDPA decode tier for PagedAttention (#32493)

hariharans29 committed 7 days ago

Verified 4725a200

Add FP8/FP4 block-quantized support to GatherBlockQuantized contrib op (#32480)

Copilot committed 7 days ago

Verified aad07ef2

Older