microsoft/onnxruntime

Pull Requests Commits

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

Return windows error code for library loading and unloading failure (#5036)

hariharans29 committed 5 years ago

Verified a9db287b

Add more huggingface models in benchmark tools (#4986)

wangyems committed 5 years ago

Verified b4e9e98c

Neg Gradient (#5022)

SherlockNoMad committed 5 years ago

Verified a935731b

Enable metadata and signature changes in graph transformers (#4783)

Dudeldu committed 5 years ago

Verified 4a0f6595

Change session option values if they don't work with EPs being registered for the session (#4991)

hariharans29 committed 5 years ago

Verified 4fd4b741

Render Operator documentation as compliant markdown (#3658)

natke committed 5 years ago

Verified 8a03b6e5

Improve performance of CUDA implementations for GatherElements and Greater, Equal and Less (#4989)

yuslepukhin committed 5 years ago

Verified e1901a7e

Build system enhancements (#5012)

Changming Sun committed 5 years ago

Verified d5d5e37e

Fix checkpoint API and improve loss scaler handling (#4950)

Thiago Crepaldi committed 5 years ago

Verified aabed34d

Fix fetches when eval_step's input is a subset of train_step's input (#4966)

Thiago Crepaldi committed 5 years ago

Verified eebc2ccc

Pass Model Path to TensorProtoToMLValue from Constant Folding for External Inputs (#5000)

Lafi7e committed 5 years ago

Verified a6e219de

Fix for Android ORT android initOsArch exception (#5006)

guoyu-wang committed 5 years ago

Verified 5651c232

Missing header for `std::once_flag` and `std::call_once`. (#5010)

xkszltl committed 5 years ago

Verified 44b3accb

Fix a warning in global_thread_pools/test_inference.cc (#4987)

Changming Sun committed 5 years ago

Verified 9902b570

Port #4920 into the new pytorch frontend (#4965)

Thiago Crepaldi committed 5 years ago

Verified f38f2d5b

Remove public default ctor in PyInferenceSession and replace it with a protected ctor (#4990)

hariharans29 committed 5 years ago

Verified d30dd41c

Remove evaluate telemetry due to redundancy (#4996)

Ryan Lai committed 5 years ago

Verified c6a3620b

Use raw attention mask in BERT related fusions (#4889)

tianleiwu committed 5 years ago

Verified a47cae03

Liqun/new api orttraining test transformers (#4982)

liqunfu committed 5 years ago

Verified d79af260

Add Cmake config for onnxruntime_NO_EXCEPTIONS (#4975)

guoyu-wang committed 5 years ago

Verified 64237d99

Newer Older