microsoft/onnxruntime

Pull Requests Commits

Fix an issue where GetLocationForNodeInput blindly put input on allocator with OrtMemTypeDefault memory type

cloudhan committed 3 years ago

166bd8af

Infer output memory type for MemcpyFromHost.

cloudhan committed 3 years ago

58601f63

Fix Reduced ops pipeline (#10861)

hariharans29 committed 3 years ago

Verified 0d8d44d0

Add additional NNAPI QDQ test cases for expected failure path (#10769)

YUNQIUGUO committed 3 years ago

Verified 7e9dfe62

Update convert_onnx_models_to_ort.py to support runtime optimizations. (#10765)

edgchen1 committed 3 years ago

Verified e53422c6

Creating test case for printing ort tensor (#10850)

ajindal1 committed 3 years ago

Verified 03181cae

Workaround https://github.com/microsoft/STL/issues/434#issuecomment-921321254

tiagoshibata committed 3 years ago

ce10d7d2

Remove kernel32 defaultlib

tiagoshibata committed 3 years ago

d1f059c8

Remove cpuinfo from WCOS builds

tiagoshibata committed 3 years ago

ddf1a312

Fix wrong percentile values returned during calibration (#10847)

mfuntowicz committed 3 years ago

Verified c4f73af2

Add support for opset 16 to transpose optimizer. (#10841)

skottmckay committed 3 years ago

Verified 6ac3b8d4

Correct a comment

Changming Sun committed 3 years ago

2d2eebb8

avoid using LocalFree on FormatMessageW buffer (#10796)

Ryan Lai committed 3 years ago

Verified 2e7592dd

add python binding for RunOptions config entry (#10694)

kota-row committed 3 years ago

Verified 64556888

don't clear grad_fns & add test (#10671)

pengwa committed 3 years ago

Verified d478a53d

Use separate build directories for full and mobile iOS packages. (#10835)

edgchen1 committed 3 years ago

Verified 1a62306d

remove unused six in code and CIs (#10832)

jcwchen committed 3 years ago

Verified 5202efd1

Patch absl so that it doesn't disable important VC++ warnings (#10836)

Changming Sun committed 3 years ago

Verified f87a06cd

Mark end of version 11 C API. (#10803)

Pranav Sharma committed 3 years ago

Verified 97ae44d0

Abjindal/torch api change gelu (#10833)

ajindal1 committed 3 years ago

Verified 3ae2bfae

Address performance issue with abseil flat_hash_table. (#10819)

yuslepukhin committed 3 years ago

Verified 1d545dfe

Fix bug in MemcpyToHost (#10816)

hariharans29 committed 3 years ago

Verified e80ff632

Detect runtime CUDA JIT and warn the user (#10781)

RyanUnderhill committed 3 years ago

Verified 9853eaa1

Uninstall onnxruntime-training before running local tests (#10827)

Changming Sun committed 3 years ago

Verified cc3a3476

Add micro-benchmarks for Attention and SkipLayerNormalization ops. (#10798)

zhangyaobit committed 3 years ago

Verified 9cbcc93e

changing gelu backward op and adding required files (#10813)

ajindal1 committed 3 years ago

Verified 1c313f44

Make QDQSelectorActionTransformer() is_int8_allowed parameter required. (#10820)

edgchen1 committed 3 years ago

Verified 0293e525

Update protobuf submodule (#10801)

Changming Sun committed 3 years ago

Verified cc6bc34c

Make training CUDA kernels to adhere established code structure patterns (#10735)

yuslepukhin committed 3 years ago

Verified 58521fb8

Making the Java tests faster by optionally disabling ones which require running multiple JVMs. (#10811)

Craigacp committed 3 years ago

Verified 4ef81b14

Older