microsoft/onnxruntime

Pull Requests Commits

try single cuda stream

chentaMS committed 3 years ago

30659444

a workaround for conv

chentaMS committed 3 years ago

9849eb98

fix the partial execution in training

chentaMS committed 3 years ago

0977c8ac

try single thread

chentaMS committed 3 years ago

9556c867

use task counter as execution complete flag

chentaMS committed 3 years ago

0d536aac

chentaMS committed 3 years ago

e9749cc4

fix training break

chentaMS committed 3 years ago

5239aac2

Fix Segfault caused by abnormal exit in PE branch (#12535)

RandySheriffH committed 3 years ago

Verified d41e337d

Merge branch 'master' into PE

jslhcl committed 3 years ago

Verified c533865d

Eager Mode - Support Concatenation via aten::cat.out (#12527)

msftlincoln committed 3 years ago

Verified 0d9a02e6

Python module for dumping activation tensors when running an ONNX model (#12474)

chenfucn committed 3 years ago

Verified 47b787c2

Load checkpoint in cpp (#12352)

AdamLouly committed 3 years ago

Verified 2681648f

Fix build break when DEBUG_NODE_INPUTS_OUTPUTS enabled (#12497)

jslhcl committed 3 years ago

Verified 3f59ea31

Add codeowners for requirement files (#12512)

faxu committed 3 years ago

Verified ee3b7574

[CUDA] SoftmaxCrossEntropy Kernels Refactor (#12482)

Lafi7e committed 3 years ago

Verified 2bed0d4a

[CUDA] Mod Op Kernel (#12499)

Lafi7e committed 3 years ago

Verified cfa09d16

Improve the compilation speed when compiling for multiple architectures. (#12490)

pengwa committed 3 years ago

Verified a2dc3e9e

Incrementally free initializers while saving to OrtValue instances (#12485)

skottmckay committed 3 years ago

Verified 56bd96a3

remove the link the comments (#12510)

HectorSVC committed 3 years ago

Verified 730240d2

[Java] JNI refactor for ONNX Tensor (#12281)

Craigacp committed 3 years ago

Verified 8a86b346

new quantized operators split (#12495)

jchen351 committed 3 years ago

Verified 8c5c2834

Fix various warning in kernel explorer (#12501)

cloudhan committed 3 years ago

Verified 9c055770

set zero point to 0 if all value are 0.0 (#12470)

yufenglee committed 3 years ago

Verified bdd6b00c

Avoid false-positive dependent name lookup error by not depending on auto keyword (#12483)

cloudhan committed 3 years ago

Verified ddea1e48

Merge branch 'PE' of https://github.com/microsoft/onnxruntime into PE

chentaMS committed 3 years ago

fb974be8

in memcpytohost, sync stream before invoke next kernel

chentaMS committed 3 years ago

f9a10a67

fix build break

Cheng Tang committed 3 years ago

129806d9

fix linux break

Cheng Tang committed 3 years ago

7f50c5b4

update the memory reuse planner; disable memory pattern for GPU pinned memory

chentaMS committed 3 years ago

3c733a8e

DML EP fix training build error (#12461)

fdwr committed 3 years ago

Verified eb90b52a

Older