microsoft/onnxruntime

Pull Requests Commits

Update GPU packaging pipelines to cuda11 and fix the other build break issues (#6585)

Changming Sun committed 5 years ago

Verified b5bd14fc

Support no bias in layernorm and skiplayernorm op (#6554)

gh-yewang committed 5 years ago

Verified 82229c8e

Support to allow user to specify compute stream per session (#3723)

weixingzhang committed 5 years ago

Verified 299ace07

OpenVino add build_shared_lib flag in the build command (#6560)

sfatimar committed 5 years ago

Verified 973c3917

Let execution fall back to CPU EP if Compile of a partition on current EP fails (#6580)

guoyu-wang committed 5 years ago

Verified 68193e28

add set_model_dir and update ONNX (#6119)

jcwchen committed 5 years ago

Verified f2ce3aae

Enable type reduction for Gather CPU kernel. (#6579)

edgchen1 committed 5 years ago

Verified 3b376da3

Add more kernels that have typed registrations to the operators we track type usage for. (#6565)

skottmckay committed 5 years ago

Verified c5d25383

Tile perf enhancements - continued (#6561)

hariharans29 committed 5 years ago

Verified f14c621c

Add type reduction support to Slice and Transpose (#6547)

skottmckay committed 5 years ago

Verified c49d1dbc

[Node.js binding] support NPM v7+ (#6559)

fs-eire committed 5 years ago

Verified 89627a81

remove keras example from python documentation (#6574)

xadupre committed 5 years ago

Verified 615acf15

Update link in readme (#6537)

prasanthpul committed 5 years ago

Verified 4e61e254

Reuse reduction_functions.cu

jessebenson committed 5 years ago

d914e29f

Pick up changes from:

jessebenson committed 5 years ago

3c441849

Fix merge conflict.

jessebenson committed 5 years ago

a9e4d70b

Fix scratch buffer early free.

jessebenson committed 5 years ago

76fcebd0

Delete ROCM-specific reduction code that is identical to CUDA reduction code.

jessebenson committed 5 years ago

86ac11af

Code formatting.

jessebenson committed 5 years ago

5d879270

Disable a couple more unsupported tests.

jessebenson committed 5 years ago

21a47ec8

Update remaining reduction ops to use MIOpen. double datatype is not supported, so disable those typed kernels.

jessebenson committed 5 years ago

0b147702

jessebenson committed 5 years ago

a28ddb85

Reuse CUDA's reduction_functions.cc

jessebenson committed 5 years ago

19613292

jessebenson committed 5 years ago

4c1db50d

Add reduce template parameters.

jessebenson committed 5 years ago

554184bc

Update reduction_all.cu

jessebenson committed 5 years ago

c4b6559b

Partial updating of ROCM reduction code.

jessebenson committed 5 years ago

5fc377f2

Cast Op performance fix. (#6509)

edgchen1 committed 5 years ago

Verified 318b82ca

Don't resolve symlink in resolve_executable_path(). (#6540)

edgchen1 committed 5 years ago

Verified 2ef792ae

Merge CPU packaging pipelines (#6480)

Changming Sun committed 5 years ago

Verified aa31ba57

Newer Older