microsoft/onnxruntime

Pull Requests Commits

commit several fixes

RandyShuai committed 5 years ago

b32e92ad

add optimized version

RandyShuai committed 5 years ago

9af05c68

Deprecate testMNISTTrainingAndTestingOpset10 (#4927)

liqunfu committed 5 years ago

Verified 84c22212

Allow sharing of initializers between sessions. (#5092)

Pranav Sharma committed 5 years ago

Verified 974b9bfc

Revert to using release SafeInt repo now that it supports a build with exceptions disabled. (#5233)

skottmckay committed 5 years ago

Verified e0719a10

Fix TransposeScaleMatMul and MatMulScaleFusion issues (#5230)

edgchen1 committed 5 years ago

Verified e9671e93

Fix a bug in EmbedLayerNorm fusion (#5150)

wangyems committed 5 years ago

Verified 65740deb

Create profile for all dynamic shape input tensors (#5229)

stevenlix committed 5 years ago

Verified aefb2cc4

Fix WinML warnings (#5228)

tiagoshibata committed 5 years ago

Verified cd663d58

[ORT Mobile] ORT Minimal E2E CI (#5200)

guoyu-wang committed 5 years ago

Verified 78a29aeb

Preserve relative order of the results and the tests. (#5225)

yuslepukhin committed 5 years ago

Verified 8ee4e822

using GPU_WARP_SIZE to make kernel portable between AMD and Nvidia GPU (#5173)

weixingzhang committed 5 years ago

Verified b49f6a5e

Fuse softmax(a + b) in case of simple broadcast (#4937)

Suffian Khan committed 5 years ago

Verified 84589c7e

Provide option to let layernorm stash mean/var as fp32 or bfloat16 (#5215)

souptc committed 5 years ago

Verified e0b49844

Refactor onnx_test_runner (#5169)

yuslepukhin committed 5 years ago

Verified a90ab125

Remove invalid install line (#5219)

RyanUnderhill committed 5 years ago

Verified 13318ab0

Amdmigraphx improvements (#5158)

scxiao committed 5 years ago

Verified a632dd2d

remove curand_generator_ related code since it is not used. (#5220)

weixingzhang committed 5 years ago

Verified f91248e0

[Python] Move symbolic_shape_infer from nuphar to tools (#5162)

KeDengMS committed 5 years ago

Verified ce3b67e0

Remove shape inference and fix save large model(>2g) issue (#5210)

YUNQIUGUO committed 5 years ago

Verified f7c1e518

Update conversion script and process to simplify creating ORT format models and a minimal build (#5217)

skottmckay committed 5 years ago

Verified c46a4803

fix _WIN32 (#5218)

jywu-msft committed 5 years ago

Verified 1b61dfaf

Fix order of returned values in quantize_weight_per_channel (#5205)

pranav-prakash committed 5 years ago

Verified f5df9625

--shm-size=1024m to fix nccl shared memory issue (#5214)

liqunfu committed 5 years ago

Verified f37e1292

[ORT Mobile] Some updates to iOS/Android build settings (#5184)

guoyu-wang committed 5 years ago

Verified 8156e0dd

NCHWc optimizer fixes for quantized models (#5203)

tracysh committed 5 years ago

Verified 86981571

Add API to allow configuration of the global thread pools. (#5199)

Pranav Sharma committed 5 years ago

Verified d5358942

Fix softmax_warp_backward math when is_log_softmax = True and register LogSoftmax CUDA kernel (#5160)

Suffian Khan committed 5 years ago

Verified e01e0b2e

Corrects doc typos and formatting (#5201)

smkarlap committed 5 years ago

Verified 584638e5

MaxPool versioning in quantization tools. (#5194)

zhanghuanrong committed 5 years ago

Verified cd0386b6

Older