strisunshinewentingwang/pytorch

Pull Requests Commits

`cuda._is_in_bad_fork`->`_C._cuda_isInBadFork` (#87317)

malfet committed 1 year ago

c8889f4e

[Dynamo] Support optimizing over any Tensor with requires_grad = True (#87141)

yanboliang committed 1 year ago

56b150ac

Symintify pad ops (#87046)

albanD committed 1 year ago

12b2f70a

Advance nightly docker to 11.6 (#86941)

atalman committed 1 year ago

c5de535b

OpInfo: Sample input cleanup (4/n) (#86324)

peterbell10 committed 1 year ago

6eeeb881

Fix compilation warning and spurious print (#87297)

albanD committed 1 year ago

c141f28b

Tweak several test serialization to store models state_dict (#87143)

malfet committed 1 year ago

4a533f12

[maskedtensor] add docs (#84887)

george-qi committed 1 year ago

cf2be34f

Revert "[primTorch] Add refs for `softmax`, `softmin`, `log_softmax` (#84956)"

pytorchmergebot committed 1 year ago

cd216135

[FSDP] Add set_state_dict_type API to setup state_dict_type without using context manager (#86243)

fegin committed 1 year ago

c08c7997

Revert "Dynamo FX graph stack traceback fix (#87136)"

pytorchmergebot committed 1 year ago

f3cc588d

[primTorch] Add refs for `softmax`, `softmin`, `log_softmax` (#84956)

nkaretnikov committed 1 year ago

c09ca93e

[allocator] disable tests that don't work for cudaMallocAsyncAllocator (#87250)

zdevito committed 1 year ago

00c91f44

[functorch] Get rid of defunct functorch/setup.py (#87235)

zou3519 committed 1 year ago

15ca6852

[functorch] add test for torch.manual_seed inside grad transform (#87233)

zou3519 committed 1 year ago

ac80da22

[allocator] Move getFreeMutex (#87237)

zdevito committed 1 year ago

f56ce8db

Dynamo FX graph stack traceback fix (#87136)

williamwen42 committed 1 year ago

89e6078b

Reenable aot tests on windows for cuda 11.7 and up (#87193)

atalman committed 1 year ago

40d0fa53

Pin ios conda dependencies (#87229)

huydhn committed 1 year ago

86a58192

[MPS] Do not dispatch empty job in `bitwise_not` (#87286)

malfet committed 1 year ago

a79e034d

fix 0d cpu tensor handling when it's the first arg (#87273)

ngimel committed 1 year ago

6775c3e1

add an API for external backends to register custom device names (#86992)

bdhirsh committed 1 year ago

fb6826bf

Clean Inductor complication cache during dynamo dashboard run (#87246)

williamwen42 committed 1 year ago

cc64863d

functionalization: skip meta reference compute for aot autograd (#87108)

bdhirsh committed 1 year ago

b3071e2e

ban .sizes() and .strides() calls in derivatives.yaml (#86611)

bdhirsh committed 1 year ago

4801397b

symintify nll loss fns (#86915) (#87095)

anjali411 committed 1 year ago

182ee879

add support for pin memory on xpu device (#86545)

leizhenyuan committed 1 year ago

c6187ea3

[complex] conv_transpose2d (#81805)

kshitij12345 committed 1 year ago

528dd051

[TorchDynamo]: fused bias for cpu convolution path (#87050)

XiaobingSuper committed 1 year ago

232fbd90

Fixed FakeTensor not calling CompositeImplicitAutograd decomps sometimes (#87252)

Chillee committed 1 year ago

5e23074f

Newer Older