microsoft/onnxruntime

Pull Requests Commits

exclude external stream

RandyShuai committed 3 years ago

77d5b14f

cancel stream-reseting of cudnn/cublas handles

RandyShuai committed 3 years ago

609e0bb4

RandyShuai committed 3 years ago

572e91de

resolve conflict

RandyShuai committed 3 years ago

e889fc91

chentaMS committed 3 years ago

daa1dd63

reset cudnn&cublus handle on context

RandyShuai committed 3 years ago

0550fa72

Chenta/partition stream with device (#12656)

souptc committed 3 years ago

Verified d271a010

chentaMS committed 3 years ago

19eb33be

Merge branch 'main' into PE

souptc committed 3 years ago

Verified 856efe3a

[oneDNN] Improve DequantizeLinear operator performance. (#12611)

Erick Muñoz committed 3 years ago

Verified 82b724fa

Add BuildError for --gen_doc and --enable_training (#12630)

Thiago Crepaldi committed 3 years ago

Verified d1ba8015

Replace to lock_guard as lighter class for locking (#12616)

yuslepukhin committed 3 years ago

Verified 9481893b

weight matching (#12607)

chenfucn committed 3 years ago

Verified f2db6bb2

Fix a build error (#12600)

hmchen-github committed 3 years ago

Verified 8a038b9b

Improve LongformerAttention performance: AddBiasTranspose and New weight format (#12448)

tianleiwu committed 3 years ago

Verified ce01ed02

remove useless files

chentaMS committed 3 years ago

747946eb

add support for in-stream sync

chentaMS committed 3 years ago

372a728b

chentaMS committed 3 years ago

83e21dfb

Refactor with std::variant (on device training) (#12383)

pengwa committed 3 years ago

Verified 7df2e8c5

try to put copy on stream

chentaMS committed 3 years ago

6f5922af

try blocking stream

chentaMS committed 3 years ago

6b5c534a

Replace references to onnxruntime 'master' with 'main' in Dockerfiles. (#12550)

edgchen1 committed 3 years ago

Verified caabfcd9

Add build option to link TensorRT prebuilt parser (#12602)

yf711 committed 3 years ago

Verified 9d10badc

[Java] JNI refactor for OrtSession (#12496)

Craigacp committed 3 years ago

Verified 733db314

run rnn ops on the specific stream

chentaMS committed 3 years ago

fbe8987f

fix the memory set async issue

chentaMS committed 3 years ago

598fc6f9

test single stream

chentaMS committed 3 years ago

71da10bb

fix training build break

chentaMS committed 3 years ago

c34ead1e

QDQ debugger - activations compare (#12544)

chenfucn committed 3 years ago

Verified eb6aa861

release calibrator before deleting temporary files (#12601)

yufenglee committed 3 years ago

Verified 30ee5a4f

Older