strisunshinewentingwang/pytorch

Pull Requests Commits

Reenable cusparse SpMM on cuda 10.2 (#42556)

malfet committed 4 years ago

3c7fccc1

handle multiple returns properly in boxing wrappers (#42437)

Basil Hosmer committed 4 years ago

78f4cff8

Reduce the output overhead of OutputColumnMaxHistogramObserver by enabling changing bin_nums, Update the observer_test.py

Yongbin Gu committed 4 years ago

d45e2d3e

Install typing_extensions in PyTorch CI (#42551)

malfet committed 4 years ago

61027a1a

[JIT] Fix torch.jit.is_tracing() (#42486)

Meghan Lele committed 4 years ago

29700c00

[ONNX] Enable lower_tuple pass for custom layer (#41548)

BowenBao committed 4 years ago

afa489de

test: Disable test_strided_grad_layout on ROCM (#42561)

seemethere committed 4 years ago

ccc831ae

Automated submodule update: FBGEMM (#42496)

facebook-github-bot committed 4 years ago

c3e2ee72

Fix the bug in THCTensor_(baddbmm) and ATen's addmm_cuda for strided views input (#42425)

IvanYashchuk committed 4 years ago

b9e68e03

Implement sort for string in aten (#42398)

gmagogsfm committed 4 years ago

317b9d3b

Fix doc build (#42559)

malfet committed 4 years ago

56fc7d03

Add private API to support tensor lists: _foreach_add(TensorList tensors, Scalar scalar) (#41554)

izdeby committed 4 years ago

e995c3d2

.circleci: Have python docs always push to site (#42552)

seemethere committed 4 years ago

a0695b34

[vulkan][asan] Fix Invalid Memory ops (#41224)

IvanKobzarev committed 4 years ago

91d87292

[vulkan] reshape op (#41223)

IvanKobzarev committed 4 years ago

0d1a6897

Clean up CUDA Sleep and Tensor Initialization in ProcessGroupNCCLTest (#42211)

osalpekar committed 4 years ago

e97e8736

TearDown function for ProcessGroupNCCLTest Initializer (#42209)

osalpekar committed 4 years ago

3ca36179

Moving ProcessGroupNCCLTest to Gtest (#42208)

osalpekar committed 4 years ago

2b8e7e2f

[TensorExpr] Properly handle all dtypes of the condition in evaluation of IfThenElse exprs. (#42495)

Mikhail Zolotukhin committed 4 years ago

b3ffebda

[TensorExpr] Properly handle all dtypes in evaluation of Intrinsics exprs. (#42494)

Mikhail Zolotukhin committed 4 years ago

c334ebf1

[TensorExpr] Properly handle all dtypes in evaluation of CompareSelect exprs. (#42493)

Mikhail Zolotukhin committed 4 years ago

38a99844

Revert "Revert D22360735: .circleci: Build docker images as part of C… (#40950)

seemethere committed 4 years ago

5939d8a3

Remove redundant kernels calling TypeDefault in VariableType codegen. (#42031)

Ailing Zhang committed 4 years ago

4b42a5b5

Initialize uninitialized variable (#42419)

wconstab committed 4 years ago

94e8676a

Fix read/write bulk data (#42504)

gunandrose4u committed 4 years ago

d2a2ac4e

fix discontiguous inputs/outputs for cummin/cummax (#42507)

ngimel committed 4 years ago

ec898b1a

Add NCCL Alltoall to PT NCCL process group (#42514)

Srinivas Sridharan committed 4 years ago

ecb88c5d

Improve the documentation of DistributedDataParallel (#42471)

0x4f5da2 committed 4 years ago

b56db305

Batching rules for: chunk, split, unbind (#42480)

zou3519 committed 4 years ago

f3e8fff0

Batching rules for: torch.movedim, torch.narrow, Tensor.unfold (#42474)

zou3519 committed 4 years ago

f1d7f001

Newer Older