strisunshinewentingwang/pytorch

Pull Requests Commits

Revert "RowwiseMoments: use float as acc type for bfloat16 inputs (#81850)"

pytorchmergebot committed 2 years ago

7e6da2fb

Add mac-mps workflow (#83296)

malfet committed 2 years ago

b6156fd4

Register nested matmul as an addition to CompositeImplicit (#82786)

YifanShenSZ committed 2 years ago

7f18ef14

Propagate CUDAOutOfMemoryError to Python. (#83146)

zdevito committed 2 years ago

41287123

[BE] Fix lint in trymerge.py (#83293)

malfet committed 2 years ago

cf6a91b7

[ROCm] Enable bf16-related tests in test_c10d_nccl.py and test_grad_layout_1devicemodule_1replicaperprocess (#82020)

hubertlu-tw committed 2 years ago

cd18b78d

Fix and unskip cpp extension tests for ARM (#83115)

soulitzer committed 2 years ago

b1896255

Set default qengine to QNNPACK on ARM for quantization tests (#83097)

soulitzer committed 2 years ago

95865132

Unskip CompositeCompliance tests for ARM (#83089)

soulitzer committed 2 years ago

ba53efa6

Autograd graphtask trim unnecessary edges (#82544)

yueyericardo committed 2 years ago

382ef1fd

Add assertions to fix torch::jit::load bugs (#79192)

apach301 committed 2 years ago

d438e867

When encountering dynamic types, one should cast it recursively. (#83218)

qihqi committed 2 years ago

2ae1afd6

Fix issues with Werror=range-loop-construct (#83273)

seemethere committed 2 years ago

e88dea0c

[quant][fx] Add more detailed docs for prepare_fx/prepare_qat_fx/convert_fx (#83132)

jerryzh168 committed 2 years ago

bce1540f

Limits constant chunk propagation for pw-node-only (#83083)

jjsjann123 committed 2 years ago

a395f6e8

Fix TORCH_CHECK macros when glog is used (#83216)

wconstab committed 2 years ago

abb2204f

Change the input of `mvit_v2_s` on the FX test (#83242)

datumbox committed 2 years ago

cff55682

Hack up make_fx to natively support varargs (#83210)

ezyang committed 2 years ago

63f35f1a

[mergebot] Update merge bot messages to explain flags (#82907)

zengk95 committed 2 years ago

23b5fcaa

[functorch] Add some more view+inplace grad+vmap tests (#83176)

zou3519 committed 2 years ago

fa54021a

[functorch] vjpvjp inplace testing (#83119)

zou3519 committed 2 years ago

f8c408b7

[functorch] in-place testing for test_vjp (#83114)

zou3519 committed 2 years ago

ffc4a502

[functorch] in-place jvp testing (#83077)

zou3519 committed 2 years ago

3dc402fd

Propose code owners from Intel (#80218)

jgong5 committed 2 years ago

beceb8b9

Add nvfuser support for prims.sign and refs.sign (#83167)

IvanYashchuk committed 2 years ago

7191ae58

[xla hash update] update the pinned xla hash (#83236)

pytorchmergebot committed 2 years ago

c35db967

CUDA trace Python hooks (#82824)

sypneiwski committed 2 years ago

916def84

Fix broken FX tests (#83187)

datumbox committed 2 years ago

6a09847c

[PyTorch] Solved two syntax issues when dumping execution graph result to json file. (#81854)

mingyul-fb committed 2 years ago

ccfbbc0f

Preserve node's stack trace during retrace (#83050)

SherlockNoMad committed 2 years ago

69156764

Newer Older