microsoft/onnxruntime

Pull Requests Commits

Missing directory

eserscor committed 48 days ago

886708da

eserscor committed 50 days ago

8ba7470a

eserscor committed 50 days ago

8f340a80

Try speedup checkout

eserscor committed 50 days ago

b214dfdd

Fix Python API docs ONNX dependency (#32811)

damdoo001-arm committed 50 days ago

Verified 3482ff20

Validate dynamic pads input access (#32654)

apsonawane committed 51 days ago

Verified f1fae4c9

Fuse BatchNormalization into ConvTranspose (#29542)

the0cp committed 51 days ago

Verified b61f007b

Enable non-causal WebGPU PagedAttention for full-context blocks (#32763)

aciddelgado committed 51 days ago

Verified 83e48928

WebGPU: add physical Dawn adapter index selection (#32319)

tianleiwu committed 51 days ago

Verified d306e3ee

Reduce coordinate work in CPU N-D Col2Im (#32800)

Hosi121 committed 51 days ago

Verified 58c827ab

Optimize FP16 LayerNorm and RMSNorm on x86 (#32715)

xadupre committed 51 days ago

Verified f403a9da

[WebGPU EP] adds state update capacity support for GatedDeltaNet (#32767)

prathikr committed 51 days ago

Verified f01b1d11

Add WebGPU MoE and expand QMoE feature support (#32447)

sayanshaw24 committed 51 days ago

Verified 4490a296

Split webgpu build into build and test. Reduce total runtime 30~45 mins. (#32768)

eserscor committed 51 days ago

Verified 8633745c

Fix WebGPU Concat with many inputs (#32783)

Copilot committed 51 days ago

Verified 096d71f0

[CPU] Fix INT2 QuantizeLinear per-axis output on the last axis (#32786)

TANGBUDU committed 51 days ago

Verified 056897df

Validate packed GQA head counts during shape inference (#32658)

danielsongmicrosoft committed 51 days ago

Verified f8751642

session: deduplicate AutoEP custom-op domains (#32711)

kuanyul-qti committed 52 days ago

Verified ae72e3ef

[WebGPU] Upgrade Dawn to the latest (#32686)

mei1127 committed 52 days ago

Verified fb8905ba

[WebGPU] Replace subgroup matrix config indices with values (#32784)

gyagp committed 52 days ago

Verified d811aa6b

Add CUDA correctness fallback for INT2 and mixed-width QMoE (#32743)

jiafatom committed 52 days ago

Verified ee5f6e7c

Remove unnecessary spectre patch (#32774)

eserscor committed 52 days ago

Verified 4119cd60

Validate protobuf consistency when ONNX comes from an installed package (#31346)

Copilot committed 52 days ago

Verified f0fd4e2e

[WebGPU] Organize self-contained provider tests (#32729)

gyagp committed 52 days ago

Verified 9228a096

[WebNN EP] Support GroupNormalization, GroupNorm and SkipGroupNorm (#32606)

Honry committed 52 days ago

Verified 65bb482b

[CUDA] Pass Concat per-input metadata by value (#32119)

tianleiwu committed 52 days ago

Verified 8cf10e7b

Add H128 INT8 paged speculative XQA (#32740)

baijumeswani committed 52 days ago

Verified 41426bd5

[CUDA] Bound fpA_intB tactic profiling time for expensive GEMMs (#32758)

tianleiwu committed 52 days ago

Verified 1378d6be

Update Android minimal size baseline (#32772)

jiafatom committed 52 days ago

Verified d7c441f0

Move the WebGPU builds back to a specialized pool (#32765)

eserscor committed 53 days ago

Verified bed64d76

Older