microsoft/onnxruntime

Pull Requests Commits

minor fix op def

cloudhan committed 1 year ago

3ccf230f

feat: allow sm80 tc task to support arbitrary gqa ratio below 8 and move gqa4 kernels to use it

cloudhan committed 1 year ago

733c0e49

feat: integrate load balanced impl

cloudhan committed 1 year ago

6b663a39

cloudhan committed 1 year ago

098417ec

fix: adapt to new interface

cloudhan committed 1 year ago

a473cbd4

feat: update integration to include tensorcore implementation

cloudhan committed 1 year ago

d4c4eac2

fix(cmake): add split compilation defs

cloudhan committed 1 year ago

a2e519a2

test: add ort integration test

cloudhan committed 1 year ago

ca3ca3db

wip: add paged attention op kernel

cloudhan committed 1 year ago

8cdf11fb

feat: add cute based paged attention impl

cloudhan committed 1 year ago

5114ac62

[JS/WebGPU] Add GatherBlockQuantized op support (#21734)

satyajandhyala committed 1 year ago

Verified af18824f

[CUDA] enable causal in MultiHeadAttention (#21852)

tianleiwu committed 1 year ago

Verified ad382120

[js/webgpu] Enable pad f16 uniform (#21691)

axinging committed 1 year ago

Verified d9c57ac7

sign native dll with correct cert (#21854)

mszhanyi committed 1 year ago

Verified 2877de73

Fix typo for react native pipeline (#21845)

carzh committed 1 year ago

Verified 983c4d57

[MIGraphX EP Support]Remove default noopt for Migraphx EP in Benchmark.py (#21843)

TedThemistokleous committed 1 year ago

Verified 9a704756

[js/webgpu] optimize MatmulNBits (#21747)

qjia7 committed 1 year ago

Verified 87165b92

Refine `op_types_to_quantize` argument handling in matmul_4bits_quantizer.py (#21815)

duanshengliu committed 1 year ago

Verified 4af62918

Replace "DML CPU" Allocator with onnxruntime::CpuAllocator (#21818)

smk2007 committed 1 year ago

Verified 44dcc3aa

[CoreML EP] Fix ArgMaxOpBuilder::AddToModelBuilderImpl() nullptr Node access. (#21797)

edgchen1 committed 1 year ago

Verified 5726318e

enable large initializer offset align for save external data in ORT (#21604)

frank-dong-ms committed 1 year ago

Verified 4c4ae1e4

[js/webgpu] Optimize conv1d by conv2d (#19388)

qjia7 committed 1 year ago

Verified 27a68905

Ovep 1.19 bug fix 2 (#21829)

preetha-intel committed 1 year ago

Verified 0368dd4e

[VitisAI] optimize model clone (#21706)

BoarQing committed 1 year ago

Verified 37a7dd7d

Revert "Upgrade emsdk from 3.1.59 to 3.1.62" (#21817)

guschmue committed 1 year ago

Verified ba7baae9

[QNN EP] Apply workaround for Conv validation bug when bias input is implicit (#21764)

adrianlizarraga committed 1 year ago

Verified 514b4699

Do not allow clearing Android logs if the emulator is not running (#21578)

jchen351 committed 1 year ago

Verified 6c1a3f85

[VSINPU]Update vsinpu patches (#21402)

chenfeiyue-cfy committed 1 year ago

Verified ff3e8b02

[VitisAI] remove wrong error msg, required by Microsoft (#21715)

BoarQing committed 1 year ago

Verified 3ff8ca29

[CUDA] Update benchmark_mha.py to capture debug info to identify sdpa kernel (#21804)

tianleiwu committed 1 year ago

Verified 25d7a4fa

Older