microsoft/onnxruntime

Pull Requests Commits

RandySheriffH committed 4 years ago

62024ddb

Support eval mode and torch.no_grad context in ORTModule and restructure ortmodule.py (#7162)

baijumeswani committed 4 years ago

Verified 844361bc

fix for using tensorrt:20.12 base image (#7264)

jywu-msft committed 4 years ago

Verified 025abf99

CUDA ConvGrad Kernel (#7227)

SherlockNoMad committed 4 years ago

Verified 4bc17ca0

Fix initializer counts when used as graph output (#7260)

TomWildenhain-Microsoft committed 4 years ago

Verified 8219518a

Remove ROCM workaround for half-to-double cast.

jessebenson committed 4 years ago

2ec452cd

Avoid passing zero bias to Gemm in gradients (#7244)

Derek Murray committed 4 years ago

Verified 25e261f1

build ONNXRuntime into WebAssembly (#6478)

fs-eire committed 4 years ago

Verified 405ca490

Not-where fusion (#7182)

ashbhandare committed 4 years ago

Verified 2aa89989

QDQ: type conversion and more ops support (#7243)

yufenglee committed 4 years ago

Verified 790fc11e

Allocate external Rocm allocator via PyBind (#7148)

raviskolli committed 4 years ago

Verified 5d759e18

Update opset for other training graphs to 12. (#7259)

Derek Murray committed 4 years ago

Verified 6308e709

Add function body to GeluGrad schema (#7190)

gramalingam committed 4 years ago

Verified a9ff4c29

Add mlas_bench tools. Starting with sconv bench and sgemm bench. (#7139)

zhanghuanrong committed 4 years ago

Verified dbcfc4be

Fix assert that the tensor's device type is 'cpu' #7248

ashari4 committed 4 years ago

Verified 56b22c1c

Add cuda kernels for GreaterOrEqual, LessOrEqual, Where; modify Clip to avoid memcpy (#7187)

ashbhandare committed 4 years ago

Verified e9ffcfa2

Update test_training_model.onnx to opset 12. (#7251)

Derek Murray committed 4 years ago

Verified c85657cf

MLAS: fix qgemm bus error with Android + ARM32 (#7250)

tracysh committed 4 years ago

Verified a9dbb511

Mem trt (#6868)

oliviajain committed 4 years ago

Verified fb40602e

Cleanup build.py (#7245)

Changming Sun committed 4 years ago

Verified 2fcd69d6

Update ContribOperators.md (#7246)

Changming Sun committed 4 years ago

Verified 5bd192c4

Make dW optional for convgrad (#7083)

pranav-prakash committed 4 years ago

Verified 3b16afc0

Update the build script for Android AAR package (#7229)

guoyu-wang committed 4 years ago

Verified c5973fbb

Add BERT-L perf regression test on MI100 and re-enable batch size test (#7240)

Suffian Khan committed 4 years ago

Verified 9f14af98

Add better model test error messaging (#7239)

Ryan Lai committed 4 years ago

Verified 10102c09

Fix Zip-Nuget-Java Packaging Pipeline (#7208)

askhade committed 4 years ago

Verified e7c5dcd5

Add detailed assertion error message (#7232)

jcwchen committed 4 years ago

Verified 3ee9b0ec

Update README.md (#7043)

mrshu committed 4 years ago

Verified 008065aa

Div mul fusion (#7183)

ashbhandare committed 4 years ago

Verified 2b851353

rename cuda_mem_limit and hip_mem_limit to gpu_mem_limit for both CUDA EP and ROCm EP (#7226)

weixingzhang committed 4 years ago

Verified 74ee24cf

Older