microsoft/onnxruntime

Pull Requests Commits

Re-enable optimizer

Pranav Sharma committed 5 years ago

0ece117e

CUDA/CPU test for NegativeLogLikelihoodLoss(12) function based loss operator. (#3793)

codemzs committed 5 years ago

Verified 4f9f6aed

bump dml version (#3792)

smk2007 committed 5 years ago

Verified b1c4d6ff

Initial checkin (#3791)

davidbrownellWork committed 5 years ago

Verified 4b8fad21

Fix tree ensemble threading bug (#3778)

skottmckay committed 5 years ago

Verified 11b81905

Add test that C is unidirectionally broadcast-able before fusing the MatMul with Add. (#3780)

skottmckay committed 5 years ago

Verified 2fc3984e

add pipeline to distributed context config (#3789)

xzhu1900 committed 5 years ago

Verified e8e95110

Function expansion support and Update ONNX to 1.7 release candidate 1. (#3782)

codemzs committed 5 years ago

Verified 517bff96

Address flaky test ReduceApiTest.Sum. (#3716)

edgchen1 committed 5 years ago

Verified 047975e4

Remove eigen device from thread pool

Changming Sun committed 5 years ago

edd5855f

fix python package linux gpu failure (#3786)

jywu-msft committed 5 years ago

Verified dcb1a215

Apply onnx-tensorrt bug fixes (#3785)

stevenlix committed 5 years ago

Verified 99ec93ea

Update documentation + Update mlas threading lib to use the new TrySimpleParallelFor. (#3779)

Pranav Sharma committed 5 years ago

Verified e42e0d47

disable cublasHgemm for training (#3769)

pengwa committed 5 years ago

Verified 29234458

collect the last few iteration latency for throuput calculation (#3766)

pengwa committed 5 years ago

Verified 98b97be6

avoid using LocalFree on FormatMessageA buffer (#3772)

Ori Levari committed 5 years ago

Verified ad63e259

Implement Pow(12) for cpu and cuda (#3727)

yuslepukhin committed 5 years ago

Verified f68a326b

Remove usage of openmp in reverse seq impl. (#3754)

Pranav Sharma committed 5 years ago

Verified 027a3649

Revert softmax kernel implementation (#3753)

Changming Sun committed 5 years ago

Verified 62c730a8

sync threads before calling next cub function (#3758)

RandySheriffH committed 5 years ago

Verified 86eaa71e

Liqun/e2e transformer test (#3540)

liqunfu committed 5 years ago

Verified af398819

Use cublasHgemm "back" for fp16 computation with Volta GPU (#3765)

pengwa committed 5 years ago

Verified 177c1357

Add Threadpool::TrySimpleParallelFor (#3759)

skottmckay committed 5 years ago

Verified 3421ec11

Add SoftmaxCrossEntropyLoss to mixed-precision-transformer. (#3760)

codemzs committed 5 years ago

Verified b9a5ed1f

Fix 'Install ONNX' CI failure (#3761)

skottmckay committed 5 years ago

Verified 9f727523

Refine GatherND CPU/CUDA Kernels & Add UTs (#3688)

pengwa committed 5 years ago

Verified 0531accc

Add Distributed Checkpointing support (#3639)

ashbhandare committed 5 years ago

Verified 58f53966

Properly creating arguments to pass to setup.py (#3744)

davidbrownellWork committed 5 years ago

Verified 7296e06d

fix warning treated as error due to ignoring return status (#3739)

suffiank committed 5 years ago

Verified ea0e2d1d

Limit the number of models run on OpenVINO (#3742)

suryasidd committed 5 years ago

Verified e529464a

Older