microsoft/onnxruntime

Pull Requests Commits

Optimize cuComputeGradInput performance. (#7479)

sabreshao committed 4 years ago

Verified e6a3308d

Fix implicit-exception-spec-mismatch warning. (#7481) (#7483)

chandru-r committed 4 years ago

Verified 6773b4f5

Fix user input order before ORTModule feed it to backend (#7456)

Thiago Crepaldi committed 4 years ago

Verified 3ee63bea

Fix some C/C++ warnings in the jni part (#7385)

Changming Sun committed 4 years ago

Verified d68cedfa

Lifhuan/force trt sequential (#7440)

lifuhuang committed 4 years ago

Verified ab373d6f

Add sequence identity for opset 14 & fix sequence insert (#7335)

BowenBao committed 4 years ago

Verified c584d482

Fix a 'Squeeze' related issue in symbolic_shape_infer.py (#7380)

thilow committed 4 years ago

Verified 22d7cde7

Fixes RelWithDebInfo build issue on windows for OV-EP (#7471)

MaajidKhan committed 4 years ago

Verified 67491520

Add function body to LayerNorm (#7378)

gramalingam committed 4 years ago

Verified 044c78f0

Add log to allow serving platforms to quantify ORT usage. (#7476)

Pranav Sharma committed 4 years ago

Verified da5c9263

Update nuphar notebook model download url (#7475)

KeDengMS committed 4 years ago

Verified 8e213292

to support multiple cuda versions in published onnxruntime-training package (#7468)

liqunfu committed 4 years ago

Verified 196e6702

Improve ConvTranspose by transposing const filter during prepacking. (#7388)

zhanghuanrong committed 4 years ago

Verified e64e30ee

Initial Objective-C API (#7366)

edgchen1 committed 4 years ago

Verified d21304ce

Add CMAKE_CUDA_ARCHITECTURES=52 to TensorRT CI pipelines (#7455)

Changming Sun committed 4 years ago

Verified 78e583d0

[wasm] fix memory info creation (#7461)

fs-eire committed 4 years ago

Verified c2418a1f

liqunfu committed 4 years ago

Verified 4cbd2cce

[JS] onnxruntime-web (#7394)

fs-eire committed 4 years ago

Verified 4ebc9c3b

NCHWc: ReorderInput improvements (#7442)

tracysh committed 4 years ago

Verified d13e5b2f

Partial graph execution perf improvements. (#7438)

codemzs committed 4 years ago

Verified 82108b18

Add pytorch version check before loading Python ONNX Runtime training module (#7377)

Thiago Crepaldi committed 4 years ago

Verified 0702a14e

Update build docker image cache cleanup build definition (#7452)

edgchen1 committed 4 years ago

Verified 4804ede5

Wait for dispatch done in RunParallelSection to fix random TP UT crash (#7443)

RandySheriffH committed 4 years ago

Verified 40568d88

Implement qlinear concat and unit test. (#7341)

zhanghuanrong committed 4 years ago

Verified ada0fbbd

Remove thread pool's cancel method and suppress some warnings (#7411)

Changming Sun committed 4 years ago

Verified b5592856

SqueezeGrad Bugfix (#7412)

Lafi7e committed 4 years ago

Verified 368e4a32

Explicitly pass cuda stream to thrust function rather than use cuda default stream implicitly (#7414)

weixingzhang committed 4 years ago

Verified ca9b3f18

dnnl matmul tensor dimension check (#7383)

jeyblu committed 4 years ago

Verified b9cbbc41

Reduce perf gap between thread pool and omp (#7333)

RandySheriffH committed 4 years ago

Verified afe912d4

Add support for ORTModule to execute the graph when ONNX drops unused… (#7424)

Thiago Crepaldi committed 4 years ago

Verified 410a81b2

Newer Older