strisunshinewentingwang/pytorch

Pull Requests Commits

Small changes/cleanup (#46950)

Elias Ellison committed 3 years ago

ad5be26b

Force LLVM Compilation for CPU Tests (#46949)

Elias Ellison committed 3 years ago

f221a19a

[NNC] Fix printing of integral doubles (#47799)

nickgg committed 3 years ago

f42cdc2e

[quant] Remove nn.quantized.ReLU module and nn.quantized.functional.relu (#47415)

jerryzh168 committed 3 years ago

1478e5ec

[NCCL] enable p2p tests (#47797)

mingzhe09088 committed 3 years ago

66f9b1de

[ONNX] Update ONNX doc for writing pytorch model (#46961)

jiafatom committed 3 years ago

9ea7a6c7

Remove references to `typing` module from setup.py (#47677)

malfet committed 3 years ago

d7c8d3cc

ATen DerivedType is dead, long live ATen RegisterDispatchKey (#47011)

ezyang committed 3 years ago

809660ff

[TorchBind] Support using lambda function as TorchBind constructor (#47819)

gmagogsfm committed 3 years ago

00a3add4

Revert "Fixed einsum compatibility/performance issues (#46398)" (#47821)

gchanan committed 3 years ago

b6cb2caa

[vulkan] Enable prepacked addmm/mm for linear layers (#47815)

SS-JIA committed 3 years ago

cfe3defd

Port bmm and baddbmm from TH to ATen (#42553)

anjali411 committed 3 years ago

e1ee3bfc

[c10d] switch ProcessGroup to be managed by intrusive_ptr (#47343)

wanchaol committed 3 years ago

553ccccc

skip test_all_reduce_sum_cuda_async test case for ROCM (#47630)

KyleCZH committed 3 years ago

859e0543

[ROCm] add skipCUDAIfRocm to test_lingalg test_norm_fro_2_equivalence_old (#47809)

jeffdaily committed 3 years ago

2df56001

Spurious numpy writable warning (#47271)

ArtistBanda committed 3 years ago

2907447c

torch.dropout: fix non-contiguous layout input (#47552)

kshitij12345 committed 3 years ago

4b25d83e

[reland][c10d] switch ProcessGroupNCCL:Options to be managed by intrusive_ptr (#47807)

wanchaol committed 3 years ago

a02baa0c

[reland] [c10d] switch Store to be managed by intrusive_ptr (#47808)

wanchaol committed 3 years ago

665ac2f7

[reland][c10d] switch ProcessGroup::Work to be managed by intrusive_ptr (#47806)

wanchaol committed 3 years ago

70ae5685

[quant] Add support for 2D indices for quantized embedding operators (#47766)

supriyar committed 3 years ago

89b371bc

[quant][graphmode][fx][refactor] insert_observer (#47782)

jerryzh168 committed 3 years ago

47386722

[quant][refactor] factor out get_combined_dict function (#47781)

jerryzh168 committed 3 years ago

dd77d5a1

add cost_aware_partition (#47673)

scottxu0730 committed 3 years ago

b46787d6

Look in named-buffers of module for tensors (#47641)

mrzzd committed 3 years ago

c5834b6a

Refactor DDP uneven inputs control flags (#47394)

rohan-varma committed 3 years ago

c9f6e70c

Workaround PyTorch debug build crash using old GCC (#47805)

malfet committed 3 years ago

e8a73fbf

Added CUDA support for complex input for torch.triangular_solve (#46916)

IvanYashchuk committed 3 years ago

52ec8b93

Free original weight after prepacking in XNNPACK based op (#46541)

Akshit Khurana committed 3 years ago

a0c4aae3

Mark overriden Tensor method `override` (#47198)

Akshit Khurana committed 3 years ago

545f624a

Newer Older