microsoft/onnxruntime

Pull Requests Commits

Create a new Dockerfile for cuda 12 and trt 8.6.1.6-1.cuda12.0 (#18000)

jchen351 committed 2 years ago

0b1e17bf

Fix Packed MultiHead Attention (#17996)

aciddelgado committed 2 years ago

c7352111

[js/web] FP16 LayerNorm, InstanceNorm, SkipLayerNorm (#17630)

dakenf committed 2 years ago

48ab0ce8

Bump @babel/traverse from 7.18.5 to 7.23.2 in /js/react_native/e2e (#17963)

dependabot[bot] committed 2 years ago

e96e3318

[CUDA] StableDiffusion XL demo with CUDA EP (#17997)

tianleiwu committed 2 years ago

4e393c8d

Fix MHA shape inference (#18009)

PatriceVignola committed 2 years ago

e4b421a9

[DML EP] Enable more MHA masks (#17882)

PatriceVignola committed 2 years ago

b7d8cdc9

support for fp16 in GetClipMinMax() (#17967)

guschmue committed 2 years ago

6122c2d3

[java] Make the backing byte buffer in an OrtValue accessible (#16578)

Craigacp committed 2 years ago

e75095df

Move a nodejs test to a different machine pool (#17970)

Changming Sun committed 2 years ago

f72e2e41

Fix AMD builds and enable testing NHWC CUDA ops in one GPU CI (#17972)

hariharans29 committed 2 years ago

41b4828a

Fix missing attribute on C# DOrtGetResizedStringTensorElementBuffer delegate (#17901)

skottmckay committed 2 years ago

b8446176

[CUDA] Fix SkipLayerNorm vectorized kernel out-of-bounds read (#17943)

tianleiwu committed 2 years ago

2b1ad48c

Bump @babel/traverse from 7.18.2 to 7.23.2 in /js/react_native (#17962)

dependabot[bot] committed 2 years ago

8518e44e

Make CUDA a NHWC EP (#17200)

gedoensmax committed 2 years ago

4e860048

[TensorRT EP] Fix cmake install (#17923)

chilo-ms committed 2 years ago

327cb4af

[js/web] optimize tsc for web: split out "npm prepare" (#17955)

fs-eire committed 2 years ago

f3ecbf10

enable training for win-wasm-ci.yml (#17954)

fs-eire committed 2 years ago

a87b5bef

add onnxruntime_float16.h into header validation list. (#17935)

mszhanyi committed 2 years ago

b908d6f2

Attempt to make the usage of the Android emulator in CIs more robust (#17903)

skottmckay committed 2 years ago

68f3eacb

Add save_attribute option to quantize_static (#17945)

zhipenghan committed 2 years ago

c48c034f

Add MatMul 4bits support on GPU (#17890)

yufenglee committed 2 years ago

74145fc5

[QNN EP] Fix index-out-of-bounds bug in Slice builder when initializer is shared (#17905)

adrianlizarraga committed 2 years ago

d10046a6

[WebNN EP] Fixed wasm heap overflow for big model (#17902)

Honry committed 2 years ago

837b62d4

Custom op shape inference API (#17737)

RandySheriffH committed 2 years ago

7b86b653

Support output cross qk, dtw and more for whisper model (#17500)

zhanghuanrong committed 2 years ago

d76afb58

Update eval_squad to use API of latest optimum (#17918)

tianleiwu committed 2 years ago

68831ecd

Fix build break with cuda 12.2 (#17922)

yufenglee committed 2 years ago

bf3bb1a1

[TensorRT EP] Add back creation cudnn and cublas using external cuda stream (#17912)

chilo-ms committed 2 years ago

1f66293f

Remove half of the compiled shaders for gridsample with unused types (#17928)

smk2007 committed 2 years ago

dab99fe3

Older