microsoft/onnxruntime

Pull Requests Commits

Deprecating Horovod and refactored Adasum computations (#5468)

Tixxx committed 5 years ago

Verified 32c67c29

Minor fix to satisfy c++14 (#6162)

Pranav Sharma committed 5 years ago

Verified efa1b0d8

Using a map of of ops to stages as input of partition function. (#5940)

Juliana Franco committed 5 years ago

Verified 36c03b32

MLAS: add NEON version of int8 depthwise convolution (#6152)

tracysh committed 5 years ago

Verified 503b61d8

Fix clean_docker_image_cache.py detection of image pushes. (#6151)

edgchen1 committed 5 years ago

Verified 0fa04bdc

Revert "work around of the build break in mac (#6069)" (#6150)

Changming Sun committed 5 years ago

Verified 344a2a8e

Fix edge case in BFCArena where allocation failures could lead to an infinite loop. (#6145)

skottmckay committed 5 years ago

Verified 72505622

Partition initial optimizer state for Zero-1 (#6093)

ashbhandare committed 5 years ago

Verified 82690486

Add gradient registration for Abs. (#6139)

Derek Murray committed 5 years ago

Verified 8fd08580

Fix TensorRT kernel conflict issue for subgraphs of control flow operators (#6115)

stevenlix committed 5 years ago

Verified aa49e476

Fix CUDA test hang: (#6138)

toothache committed 5 years ago

Verified 0978d2bf

nnapi add min max support (#6117)

guoyu-wang committed 5 years ago

Verified b648bf64

Enable running the mnist_training sample without cuda (#6085)

georgen117 committed 5 years ago

Verified 939cc9b4

Unify IExecutionProvider and IExecutionProviderFactory interfaces (#6108)

RyanUnderhill committed 5 years ago

Verified ac62cf80

Model Fusion For Bart (#6105)

liuziyue committed 5 years ago

Verified 980a93c1

remove dnnl_dll_path from post build copy (#6142)

jywu-msft committed 5 years ago

Verified 297c8248

Deprecate Python global configuration functions [Part 1] (#5923)

edgchen1 committed 5 years ago

Verified 64709b13

Minor changes to AMD element-wise kernels to converge with CUDA element-wise kernels.

jessebenson committed 5 years ago

a8d549e1

Don't mark issues that are marked as enhancement as stale (#6134)

Pranav Sharma committed 5 years ago

Verified a9548283

Reduce amount of compiled CUDA device code (#6118)

edgchen1 committed 5 years ago

Verified 9810b9e0

Enable C# .NET5 for WinML (#6120)

smk2007 committed 5 years ago

Verified a6a23db1

Unify activation and initializer alignment value (#6109)

SherlockNoMad committed 5 years ago

Verified eb5c1f0f

Liqun/move nightly pl to linux multi gpu v100 (#6024)

liqunfu committed 5 years ago

Verified cde723a1

state_dict and load_state_dict for ORTTrainer (#6095)

baijumeswani committed 5 years ago

Verified dd2e5a1a

Bump ini from 1.3.5 to 1.3.8 in /nodejs

dependabot[bot] committed 5 years ago

d4dddd99

Honor auto_pad attribute in ConvTranspose (#4271)

hariharans29 committed 5 years ago

Verified c755ca0b

Fix multi-tensor LAMB reduction to be deterministic (#6028)

Suffian Khan committed 5 years ago

Verified 6cb5d3ac

Fix DEBUG_NODE_INPUTS_OUTPUTS test by putting it in a separate process, clean up unused test_main.cc files. (#5949)

edgchen1 committed 5 years ago

Verified c8ac34d6

Introduce VariadicAlias, remove hardcoded alias limits (#6106)

SherlockNoMad committed 5 years ago

Verified a53f4dd3

Make ROCM and CUDA reduction_all code more similar.

jessebenson committed 5 years ago

38c49c24

Newer Older