microsoft/onnxruntime

Pull Requests Commits

Merge branch 'main' into aciddelgado/gqa_memeff_v2

aciddelgado committed 2 years ago

f08495f7

[js/webgpu] Optimize softmax by vector (#18153)

qjia7 committed 2 years ago

Verified 785e2b1e

optimize SLN with large dimension (#18138)

yufenglee committed 2 years ago

Verified 90d1f537

[DML EP] Handle non-raw data in dynamic graph compilation (#18160)

PatriceVignola committed 2 years ago

Verified 348a9632

Augment blockwise quantization (#18101)

chenfucn committed 2 years ago

Verified 4819fbf3

[QNN EP] Disable early termination in GetCapability (#18140)

HectorSVC committed 2 years ago

Verified be2f72a3

[js/web] fix a few package consuming problems (#18109)

fs-eire committed 2 years ago

Verified 9bba9908

Revert "Disable dml stage in windows GPU pipeline temporarily. (#18034)" (#18150)

mszhanyi committed 2 years ago

Verified 436056dc

Fix regression in perf test runner (#18139)

hariharans29 committed 2 years ago

Verified 8ebdd3bb

create memory descriptors based on the tensor dimensions (#15848)

snadampal committed 2 years ago

Verified 0e341004

Distributed Expand (#18126)

wschin committed 2 years ago

Verified 24f9c1af

Tune min version supporint custom op ComputeV2 (#18134)

RandySheriffH committed 2 years ago

Verified 8daabf3f

undo clang on unrelated files

aciddelgado committed 2 years ago

44ca857f

[WebNN EP] Remove Conv initializer constraint for GPU (#18129)

zesongw committed 2 years ago

Verified d9695dea

clang and fix test file

aciddelgado committed 2 years ago

bc1cf0aa

max sequence length for memory efficient attention

aciddelgado committed 2 years ago

e7a50ee9

Fix stale bot issue (#18064)

sophies927 committed 2 years ago

Verified 28ad3ff7

Enable global TRT timing cache (#17865)

gedoensmax committed 2 years ago

Verified 2eeafc37

Replace Transpose with Replace if they are equivalent (#18096)

guyang3532 committed 2 years ago

Verified 58f1d15d

clarify input and output formats memory efficient attention

aciddelgado committed 2 years ago

db307f36

GemmFloat8 as a contrib ops (#16051)

xadupre committed 2 years ago

Verified b5f242e9

Update python cryptography version to 41.0.4 (#18056)

xadupre committed 2 years ago

Verified c10b83eb

Distributed Reshape Implementation (#18068)

wschin committed 2 years ago

Verified 9c323106

Add updates to LLaMA scripts (#18076)

kunal-vaishnavi committed 2 years ago

Verified b79ea748

[FIX] reorder initializer (#18097)

mindest committed 2 years ago

Verified 0f3a067d

[ORTModule] ATen Efficient Attention and Triton Flash Attention (#17959)

Lafi7e committed 2 years ago

Verified b7408f73

enable reduce ops on opset18 (#18053)

souptc committed 2 years ago

Verified 37873be8

[js/webgpu] Change timestamp-query-in-passes to timestamp-query (#18108)

Yang Gu committed 2 years ago

Verified 52f49683

[CUDA] Remove shape warnings in NHWC <> NCHW unit tests (#17992)

gedoensmax committed 2 years ago

Verified b7bee621

Change cuda image for tensorRT to the one with cudnn8 (#18102)

jchen351 committed 2 years ago

Verified 7c18c60b

Newer Older