microsoft/onnxruntime

Pull Requests Commits

Fusion to support extra add after QK Matmul without mask

apsonawane committed 2 years ago

e207f082

fix fused relu activation (#18303)

guschmue committed 2 years ago

Verified 25fbc2b0

Fix rust compile issues and add GH action to run build validations and tests (#18346)

devigned committed 2 years ago

Verified 2c22b498

[Migraphx EP] Static int8 QDQ support (#17931)

TedThemistokleous committed 2 years ago

Verified 8d503138

[QNN EP] Enable option to set QNN context priority (#18315)

HectorSVC committed 2 years ago

Verified 55c19d6a

add bfloat16 support for CUDA Neg kernel (#18306)

prathikr committed 2 years ago

Verified 7a3da452

Add FlattenAndUnpad Op (#17845)

guyang3532 committed 2 years ago

Verified 4dc63692

Add tool to fix lines > 120 chars. (#18293)

skottmckay committed 2 years ago

Verified 885bf356

Bump linter versions (#18341)

justinchuby committed 2 years ago

Verified c2505407

Add a build validation for Linux ARM64 cross-compile (#18200)

Changming Sun committed 2 years ago

Verified 81253259

Update stale.yml (#18304)

sophies927 committed 2 years ago

Verified 68fab24c

Update Abseil raw_flat_hash visualization (#18329)

yuslepukhin committed 2 years ago

Verified a37e6a50

[QNN EP] Session option for graph optimization (#18262)

adrianlizarraga committed 2 years ago

Verified a0eeeafa

Add LLaMA GQA ragged batching (#18337)

kunal-vaishnavi committed 2 years ago

Verified c8def0cc

bfloat16 support for quickgelugrad (#18336)

prathikr committed 2 years ago

Verified 34f77eaa

Tune ORTModule logging experience a bit (#18298)

pengwa committed 2 years ago

Verified 2151c79b

SDXL: Update demo with dynamic shape serving with CUDA EP (#18340)

tianleiwu committed 2 years ago

Verified 8044e5f6

GQA Flash Attention with Attention Mask (#18283)

aciddelgado committed 2 years ago

Verified 3dece27f

[js] fix linter out-of-memory issue (#18307)

fs-eire committed 2 years ago

Verified 10df847b

fix typo (node)->(browser) in linux-wasm-ci.yml (#18309)

fs-eire committed 2 years ago

Verified d117a801

Do not run AOT function inlining when the model does not define any local functions (#18302)

yuslepukhin committed 2 years ago

Verified 096307c6

[js/webgpu] Simplify the Resize shader when noScale is true (#18321)

qjia7 committed 2 years ago

Verified 606356d0

implement gridsample 20 (#17744)

liqunfu committed 2 years ago

Verified 6127dd1d

add bfloat16 support for ConcatTraining and SplitTraining ops (#18280)

prathikr committed 2 years ago

Verified 83c02753

[JS/Web] Added Uniforms support to binary ops. (#18260)

satyajandhyala committed 2 years ago

Verified a16d5283

[DML EP] Add RotaryEmbedding (#18158)

PatriceVignola committed 2 years ago

Verified 800ae774

[Fix] Stages to Run couldn't be selected (#18310)

mszhanyi committed 2 years ago

Verified 9868a713

Customize _get_tensor_rank for model export in stage3 (#18294)

pengwa committed 2 years ago

Verified 4f15b427

Zhijxu/improve ortmodule python perf a little bit (#13716)

zhijxu-MS committed 2 years ago

Verified 630c877b

SkipGroupNorm fusion and SDXL Pipeline Update (#18273)

tianleiwu committed 2 years ago

Verified 00c2bf39

Older