strisunshinewentingwang/pytorch

Pull Requests Commits

Min-cut partitioner always saves tensors that are returned as-is in backward (#114970)

soulitzer committed 223 days ago

55483fc2

Update to TorchFix 0.4.0 (#119424)

kit1980 committed 223 days ago

bd9db6a9

Add cherry-pick workflow (#119352)

huydhn committed 223 days ago

5acd1f0f

[export] suppress type error (#119720)

suo committed 223 days ago

f15b5170

Restore OpInfo/ModuleInfo tests in Inductor-wrapped tests (#119693)

zou3519 committed 223 days ago

b3df3e4e

Add missing include for internal build (#119721)

albanD committed 223 days ago

e0cabeba

[inductor] Update JIT Inductor cpp wrapper entry function signature (#119280)

desertfire committed 223 days ago

70c93c60

make flash_attn_bw impl correct w.r.t. meta when k and v have different strides (#119500)

bdhirsh committed 223 days ago

02b60e76

Remove redundant CMake NUMA code (#119650)

cyyever committed 223 days ago

10789ccd

Revert "Enable x86 CPU vectorization on windows (#118980)"

pytorchmergebot committed 223 days ago

34a61c52

[DeviceIndex][3/N] Use DeviceIndex in more places (#119635)

cyyever committed 223 days ago

10f3abc6

Correctly formatting the example in get_state_dict (#119532)

jmarintur committed 223 days ago

064b6100

[ez] Add try catch for deleting old branches (#119696)

clee2000 committed 223 days ago

ad217d42

Remove unnecessary skipIfTorchDynamo from test_jit_fuser_te (#118728)

zou3519 committed 223 days ago

7eecbf8a

Update documentation for set_flush_denormal support on ARM (#119354)

maajidkhann committed 223 days ago

28c30f29

Revert "Enable fake tensor caching in fbcode by default (#118555)"

pytorchmergebot committed 223 days ago

7d780ff8

Check QNNPACK support for the platform before running test (#119139)

spzala committed 223 days ago

110919c9

Add Python 3.12 as experimental to release 2.2 (#119705)

atalman committed 223 days ago

7adfeba4

[export] improve FakeTensor serialization (#119531)

suo committed 223 days ago

82248f0b

Refactor out shape test into InputMetadata::maybe_reduce (#119559)

ezyang committed 223 days ago

482345d7

Revert "Optimize multi_tensor_apply (#119153)"

pytorchmergebot committed 223 days ago

c24b74ef

[MPS][EZ] Fix cfloat->chalf conversion on MacOS13 (#119681)

malfet committed 223 days ago

8d8fb978

fix is_ and is_not (#118978)

laithsakka committed 223 days ago

eb0f9efd

[Dynamo] Minor cleanup of redundant function lookup logics (#119666)

yanboliang committed 223 days ago

0e5b6594

Fixed hash issue in `fx_graph_cse` (#119567)

vfdev-5 committed 223 days ago

ed20e911

[reland] Fix estimate_nccl_collective_runtime (#118986)

yifuwang committed 223 days ago

27ffede8

[c10d] PGNCCL refactor part 2: Simplify ProcessGroupNCCL into single-device style (#119421)

kwen2501 committed 223 days ago

b2043c05

[c10d] explicitly abort communicators in destroy_process_group call (#119250)

shuqiangzhang committed 223 days ago

893dcac0

Clarify that legacy cat behavior only applies for 1-D tensor (#119684)

ezyang committed 223 days ago

31f00b01

Separate clang lint? (#119575)

clee2000 committed 223 days ago

059bf1ba

Newer Older