microsoft/onnxruntime

Pull Requests Commits

fix cast opset13 and other missing part of bfloat16

chentaMS committed 5 years ago

7a02aad1

chentaMS committed 5 years ago

15ae3101

add bfloat16 to transpose matmul

chentaMS committed 5 years ago

f86e764f

chentaMS committed 5 years ago

b89da45d

more fix to bfloat

chentaMS committed 5 years ago

0f4427af

bfloat16 support fix

chentaMS committed 5 years ago

5bec4927

temp disable gpu fusing transformers

chentaMS committed 5 years ago

b627bc90

Merge branch 'master' into chenta/bfloat16

chentaMS committed 5 years ago

a176c299

Simple support of MatMul U8S8 on ARM to pass tests (#4392)

yufenglee committed 5 years ago

Verified 473cd554

deprecate frontend layernorm postpass (#4372)

BowenBao committed 5 years ago

Verified 7ec9a732

Support builds without RTTI (#4333)

tiagoshibata committed 5 years ago

Verified 7fea332f

Liqun/backprop deterministic graph (#4315)

liqunfu committed 5 years ago

Verified 5dcb9b48

qlinaradd for arm/sse2/avx2 using intrinsic, enable binary broadcasting parallel (#4216)

zhanghuanrong committed 5 years ago

Verified 94c98aa0

Change the way java home is set on Mac OS for CI and Java publishing pipeline (#4385)

yuslepukhin committed 5 years ago

Verified 49268c42

BiasDropoutFusion (#4167)

SherlockNoMad committed 5 years ago

Verified 63657609

Update function body initialization for ONNX functions (#4332)

askhade committed 5 years ago

Verified 0404763f

Match More EmbedLayerNormalization Patterns for Bert Model Graph Fusion (#4354)

liuziyue committed 5 years ago

Verified 37b624b6

NCHWc + Sigmoid optimization (#4360)

tracysh committed 5 years ago

Verified 75567554

adjust bs size (#4375)

ytaous committed 5 years ago

Verified 4380b8ba

fix output shape calc for matmul (#4362)

askhade committed 5 years ago

Verified 89c6da99

Add stale bot (#4323)

faxu committed 5 years ago

Verified a4127fc1

Update Attention op to support attention mask for GPT-2 (#4330)

tianleiwu committed 5 years ago

Verified 55f25a4b

Support to build CUDA EP for NV Ampere GPU (#4345)

weixingzhang committed 5 years ago

Verified 2601f8e1

Misc fixes to Conv and ConvTranspose CUDA kernels (#4281)

hariharans29 committed 5 years ago

Verified 465140b3

Ignore one failed test in DML (#4366)

Changming Sun committed 5 years ago

Verified 35a048ef

Add nhwc support for NNAPI EP, add concat op, handle concurrent calls to NNAPI model (#4356)

guoyu-wang committed 5 years ago

Verified 5f4e63ed

Make DML operator registration constexpr (#4219)

tiagoshibata committed 5 years ago

Verified 88402f52

Minor optimization in CUDA Reduction ops (#4353)

hariharans29 committed 5 years ago

Verified 012aaa64

Cleanup SessionState. Move allocator lookup to SessionState. (#4194)

skottmckay committed 5 years ago

Verified 274e6b41

[OpenVINO] Upgrade OpenVINO docker base to Ubuntu 18.04 (#4346)

smkarlap committed 5 years ago

Verified 4a1ecd98

Older