strisunshinewentingwang/pytorch

Pull Requests Commits

clee2000 committed 2 years ago

e82e36f7

[DTensor] Fix DeviceMesh (#96861)

shintaro-iwasaki committed 2 years ago

397fb276

Cache the transformers model used in ONNX test (#96793)

huydhn committed 2 years ago

6718e3ca

Back out "Fixing a bug where allocating a 4GB block results in using 8GB of memory (#95827)" (#96796)

jaewonlee-fb committed 2 years ago

aeb3db8a

Enable TestTorchbind on Windows (#96507)

mantaionut committed 2 years ago

0eb9e01c

[MPS] LSTM grad_y missing fix (#96601)

alexdremov committed 2 years ago

62eb7a2e

Turn off split reductions for dynamic shapes (#96850)

ezyang committed 2 years ago

b249b44b

[primTorch] handle out in `sort` meta function (#96719)

nkaretnikov committed 2 years ago

bf08d138

[CI] Revert https://github.com/pytorch/pytorch/pull/96195 (#96897)

desertfire committed 2 years ago

577d930c

Add xpu device type to torch dispatch overload (#96849)

guangyey committed 2 years ago

8187c0de

Back out "Add _int_mm to expose cuBLAS int8@int8 -> int32 matmul (#94339)" (#96885)

cpuhrsch committed 2 years ago

0a53c962

fix random output issue on index_select when src is scalar and index is empty (#96408)

mingfeima committed 2 years ago

06054d7d

[TP][DTensor Op] Enable Embedding op for DTensor (#96702)

fduwjj committed 2 years ago

3405ac8a

fix flop formulas for sdpa (#96690)

Chillee committed 2 years ago

44c9ecad

Remove pytest block list (#96698)

clee2000 committed 2 years ago

8c2341c1

[memory debugging] Extract frame information from inductor (#95753)

zdevito committed 2 years ago

3162f717

Revert "Revert "[memory profiling] add a facility to gather combined C++/Python/TorchScript stack traces. (#95541)"" (#96878)

zdevito committed 2 years ago

e74f70d2

[vision hash update] update the pinned vision hash (#96906)

pytorchmergebot committed 2 years ago

1f340df3

Default specialize_int to False (#96624)

ezyang committed 2 years ago

3606f593

[FSDP] Rename to _get_orig_buffer_dtypes (#96790)

rohan-varma committed 2 years ago

308a58eb

[DDP] API to get data parallel parameters (#95097)

rohan-varma committed 2 years ago

71adb32d

Add environment variable to force flattening of 3D input tensor (#96761)

kwen2501 committed 2 years ago

3ce9aac7

Updates to the release notes scripts and documentation (#94560)

drisspg committed 2 years ago

dcafe3f2

Fix periodic job by excluding check_graph_breaks (try 2) (#96803)

wconstab committed 2 years ago

731bb6e6

fix aminmax output resize issue when input is a zero dimension tensor (#96171)

mingfeima committed 2 years ago

6d62134f

Remove un-used list. And disable pytest for public binding test. (#96684)

albanD committed 2 years ago

7c525823

Expose API to specify custom context manager for checkpoint (#96783)

soulitzer committed 2 years ago

f3db2a63

Add exceptionhandler to more distributed_c10d APIs (#96770)

H-Huang committed 2 years ago

ac7329b3

[CUDA] Use accumulate type to improve accuracy of grid_sample on half precision inputs [v2] (#96586)

xwang233 committed 2 years ago

1716709d

Output peak memory stats from dynamo torchbench perf CI (#95666)

wconstab committed 2 years ago

54cd4a67

Older