microsoft/onnxruntime

Pull Requests Commits

hariharans29 committed 5 years ago

bbc379ec

Add sequence and map support in ORT mobile file format, add UT (#5066)

guoyu-wang committed 5 years ago

Verified d922cb10

Liqun/transformer test and e2e golden numbers (#5064)

liqunfu committed 5 years ago

Verified de58720a

Register OpSet13 CUDA Kernels for BERT/UniLMv2 (#4856)

Lafi7e committed 5 years ago

Verified 84de14a8

Add a docker file for CI build CUDA 10.2 (#5065)

Changming Sun committed 5 years ago

Verified 370d194d

Implement QLinearMul in mlas (#4593)

zhanghuanrong committed 5 years ago

Verified ec88f14a

Last major set of ORT format model changes (#5056)

skottmckay committed 5 years ago

Verified b5c2932a

Parallelizing elementwise kernels (#4577)

duli2012 committed 5 years ago

Verified 6134994d

Add SetLanguageProjection C Api and use it in four projections (#5023)

zhangxiang1993 committed 5 years ago

Verified 0dad79b4

Fix initializer name only when wrapper is applied (#4920)

BowenBao committed 5 years ago

Verified 6dd4af39

Remove Cuda dependency from TensorRT shared provider (#5014)

RyanUnderhill committed 5 years ago

Verified d792af77

optimize resize op for NN mode for some fasterrcnn model (#4825)

zhanghuanrong committed 5 years ago

Verified 78bb5338

Implement QLinearSigmoid. (#5015)

zhanghuanrong committed 5 years ago

Verified 8289981f

Re-enable CI tests for the new PyTorch frontend (#5017)

Thiago Crepaldi committed 5 years ago

Verified 0fc9c504

ACL v20.02 (#4981)

Andrews548 committed 5 years ago

Verified bd215b79

Fix contrib ops unregister to match pytorch behavior (#5052)

BowenBao committed 5 years ago

Verified 73456f10

Add nodejs samples README (#5005)

natke committed 5 years ago

Verified d7502eff

Imported protoc cannot have compile options. (#5030)

xkszltl committed 5 years ago

Verified 4b9b5b61

Add packages from training docker to cgmanifest. (#5033)

Sergii Dymchenko committed 5 years ago

Verified d7984fe6

to allow parallel training with mpi4py (#4942)

liqunfu committed 5 years ago

Verified bb13b522

Add warning to non pickable models (#5037)

Thiago Crepaldi committed 5 years ago

Verified 9388d49c

Update CODEOWNERS and minor docstring fix (#5002)

Thiago Crepaldi committed 5 years ago

Verified 9d1bdef1

Add deterministic path for AllReduceL2 (used to compute gradient norm) (#5027)

Suffian Khan committed 5 years ago

Verified 546965c2

fix py packaging pipeline (#5038)

askhade committed 5 years ago

Verified 9ba2cfb7

Add flag to _internal_use to control export of contrib ops in ort trainer (#4968)

BowenBao committed 5 years ago

Verified 22ba266b

Changes to enable saving and loading an ORT format model (#4995)

skottmckay committed 5 years ago

Verified 28445c88

Remove SchedulingParams variants of ThreadPool::TryParallelFor (#5050)

tlh20 committed 5 years ago

Verified bbb9d92a

Temporarily switch SafeInt to a fork for an option to disable exceptions (#5041)

guoyu-wang committed 5 years ago

Verified fde7a2c8

Fix allocator bug (#5042)

RyanUnderhill committed 5 years ago

Verified e0d1cf19

Enable TF32 for training on A100 (#4914)

weixingzhang committed 5 years ago

Verified 32687176

Older