strisunshinewentingwang/pytorch

Pull Requests Commits

[codev] Make backport work with flatbuffer models (#88127)

qihqi committed 2 years ago

5c3666cb

Add ability to freeze storages inside functionalization (#88141)

ezyang committed 2 years ago

bb7e6254

Inline Alias into FunctionalStorageImpl (#88140)

ezyang committed 2 years ago

61f955dd

always realize output regardless of the number of reads (#88046)

Natalia Gimelshein committed 2 years ago

73c9911f

Fix meta for aten.fill, constant_pad_nd, _adaptive_avg_pool2d (#88069)

SherlockNoMad committed 2 years ago

c368c0fa

Change dynamo/distributed tests to use cuda/nccl (#88133)

wconstab committed 2 years ago

82a9de16

[BE]fix DDP when the number of output features is zero (#87793)

zhaojuanmao committed 2 years ago

44f8efd5

[9/N] [Dispatchable Collectives] Update reduce_scatter with CPU / CUDA implementations (#86166)

H-Huang committed 2 years ago

20d849b9

Reenable assert sanity testing with ADInplaceOrView reenable (#88102)

ezyang committed 2 years ago

1e5d33b6

[Reland][ONNX] Move all torch.onnx.export related tests to test/onnx (#87292)

titaiwangms committed 2 years ago

bdb14238

Update _distributed_c10d.pyi (#88088)

yhcharles committed 2 years ago

62988e4f

[FSDP()][13/N] Refactor unshard/reshard/grads (#87926)

awgu committed 2 years ago

b1750d04

[FSDP()][12/N] Easy cleanup (#87925)

awgu committed 2 years ago

8039317c

[FSDP()][10/N][11/N] Introduce composable (ctor only) (#87924)

awgu committed 2 years ago

c1e28731

[FSDP()][9/N] Refactor ctor (continued) (#87923)

awgu committed 2 years ago

78170701

[Static Runtime] Fix ReplaceWithMaybeCopy test in OSS (#88099)

Mike Iovine committed 2 years ago

23fe6c8c

Fix monitoring script for macos (#88159)

huydhn committed 2 years ago

7c6fe21a

Cleaned up the nvFuser Python Frontend Batch Norm printing (#88057)

kevinstephano committed 2 years ago

323c646c

[BE] Use default constructor in `LoggerVoidify` (#88054)

malfet committed 2 years ago

a6acbad5

call contiguous on BMM inputs for NT on CUDA (#88108)

drisspg committed 2 years ago

560786ac

Remove "prims_nvfuser" backend for TorchDynamo (#88083)

IvanYashchuk committed 2 years ago

0eea05b1

[torch::deploy] add gpu unit tests to CI (#88107)

PaliC committed 2 years ago

a8aaee77

Add support for neg to NestedTensor (#88131)

cpuhrsch committed 2 years ago

6a75a0d1

Add labeler with cpu, mkldnn, amp, NNC and quantization paths to start (#87690)

yanbing-j committed 2 years ago

708c050a

[xnnpack][lite-int][4/n] introduce serialization to delegate (#87908)

mcr229 committed 2 years ago

3aa7a528

[xnnpack][lite-int][3/n] flatbuffer serializer class (#87907)

mcr229 committed 2 years ago

8287c1d9

[xnnpack]lite-int][2/n] flatbuffer xnn_value schema (#87906)

mcr229 committed 2 years ago

7bf819b1

[xnnpack][lite-int][1/n] flatbuffer buck rules (#87826)

mcr229 committed 2 years ago

905d532d

[xnnpack][lite-int][graph-build] torchscript -> xnnpack graph (#87824)

mcr229 committed 2 years ago

aa1f9a1b

Also skip large models for normal --accuracy runs (#88086)

ezyang committed 2 years ago

d596b048

Newer Older