strisunshinewentingwang/pytorch

Pull Requests Commits

[inductor][triton] if device is a torch.device, then make cuda_properties index it correctly

soumith committed 1 year ago

b6432340

Revert "[BE] Get rid of deprecation warnings in workflows (#87152)"

pytorchmergebot committed 1 year ago

11915b31

[triton] allow cuda properties to be queried from workers (#87101)

zdevito committed 1 year ago

d36c284d

[BE] Get rid of deprecation warnings in workflows (#87152)

malfet committed 1 year ago

9da032ec

Revert "[BE] Get rid of deprecation warnings in workflows (#87152)"

pytorchmergebot committed 1 year ago

66658e1d

[Inductor] Lift the maximum depth of the Python interpreter stack to adapt large/deep models (#87130)

yanboliang committed 1 year ago

8ca7820e

[BE] Get rid of deprecation warnings in workflows (#87152)

malfet committed 1 year ago

acaf484f

Enable sdp_forward for NestedTensors (#86720)

drisspg committed 1 year ago

5fb68718

Use conda-forge in mac mps test (#87155)

huydhn committed 1 year ago

74138a8d

[vulkan] Use 2D texture types for convolution weights and biases (#86972)

SS-JIA committed 1 year ago

9d1a8edc

[vulkan] Enable 2D texture types (#86971)

SS-JIA committed 1 year ago

5b588036

Check all CUDA API calls in aten/src/ATen/test for errors (#74919) (#83556)

r-barnes committed 1 year ago

a7ed398c

Install blas from conda-forge (#87150)

huydhn committed 1 year ago

f02f0e3a

Small update to Module note (#87142)

albanD committed 1 year ago

9db7270e

Enable UBSAN mode for test_jit (#85735)

mehtanirav committed 1 year ago

fb614b18

[ci] put more logs in a folded group (#86138)

clee2000 committed 1 year ago

18cc00d3

remove dynamo hash updates (#87092)

clee2000 committed 1 year ago

e3b84f6c

Don't only apply DDP optimizer on forward frames (#87097)

davidberard98 committed 1 year ago

4fd98dfe

Add venv to gitignore (#86702)

justinchuby committed 1 year ago

09d72091

[DataPipe] Fixing interface generation in setup.py (#87081)

NivekT committed 1 year ago

0cb273b5

[ci] fix bot comment (#87127)

suo committed 1 year ago

f5ee2d88

[Docs] Remove outdated comment for sparse all-reduce (#87018)

awgu committed 1 year ago

f552eee4

handle libomp update on circleci (#86979)

clee2000 committed 1 year ago

d023e839

Use 12xlarge for nightly cpp doc generation job (#86859)

huydhn committed 1 year ago

5acf6e0e

[dynamo] Introduce `get_real_value` API to TensorVariable (#87091)

suo committed 1 year ago

48142707

[docs] Fix ScalarTensor __repr__ in Extending PyTorch example (#86330)

jmargeta committed 1 year ago

e85dbcc9

[Dynamo] More robust pyop support, module properties as args (#87020)

voznesenskym committed 1 year ago

b8007742

[ONNX] Ignore print(Tensor) during tracing (#86223)

Thiago Crepaldi committed 1 year ago

1167949b

Workarounds for cudnn_batch_norm with TorchRefsNvfuserCapabilityMode (#86796)

IvanYashchuk committed 1 year ago

31931515

add XLA backend into tensor type strings (#86881)

holimion committed 1 year ago

33343def

Older