strisunshinewentingwang/pytorch

Pull Requests Commits

Remove native_functions.yaml dependency from IndexKernel.{cpp,cu} (#66914)

peterbell10 committed 2 years ago

174eea8a

Remove native_functions.yaml dependency from TensorModeKernel.cu (#66913)

peterbell10 committed 2 years ago

f7d59894

[CUDA Pinned Memory] Alternative implementation of pinned memory allocator focusing on multi-threaded scalability (#68906)

Andrew Tulloch committed 2 years ago

ec1339a4

[LTC] Upstream some utilities (#69046)

Jiewen Tan committed 2 years ago

0cdeb586

Sparse: Implement simple unary ufuncs operators (#68887)

peterbell10 committed 2 years ago

fbaa19a6

[TensorExpr] Supress TracerWarnings in test_unsupported in test_jit_fuser_te.py. (#68757)

Mikhail Zolotukhin committed 2 years ago

3186d369

[TensorExpr] Allow for 'keepdim' argument in aten::mean in NNC's external call. (#68756)

Mikhail Zolotukhin committed 2 years ago

75ce0406

[TensorExpr] IRPrinter: print sizes and name when visiting a Buf. (#68755)

Mikhail Zolotukhin committed 2 years ago

a93f505e

[LTC] Further refactor bridge's device management (#69062)

alanwaketan committed 2 years ago

Verified 9c0ac6c9

Add option to get input dtype from user (#68751)

priyaramani committed 2 years ago

8cc9ec2f

Clean up some THC includes (#69024)

t-vi committed 2 years ago

ac1fe91d

Merging the implementations of ClearProfiling (#67575)

Gamrix committed 2 years ago

ce53baf5

cpu_kernel_vec: Hoist stride checks out of loop (#68962)

peterbell10 committed 2 years ago

e6a8d15a

Fix device type / dtype handling for parametrized test names (#65217)

jbschlosser committed 2 years ago

61ea2fc3

Fixed links to RNN docs in comments (#68828)

dozed committed 2 years ago

933d5b56

Fix typo in AdaptiveLogSoftmaxWithLoss docs (#68926)

stas00 committed 2 years ago

863f321c

Remove autograd-enabled collective APIs from distributed docs (#69011)

mrshenli committed 2 years ago

b8c36932

Vectorized: Use inline namespace instead of anonymous (#67655)

peterbell10 committed 2 years ago

17801045

[PyTorch] Change from unique_ptr to optional for RecordFunction state (#68397)

swolchok committed 2 years ago

1d041639

[PyTorch] Optimzize mergeRunCallbacks for RecordFunction (#68387)

swolchok committed 2 years ago

7194faed

Adding Linux cuda 11.5 workflows (#68745)

atalman committed 2 years ago

f1a3512b

[FX][docs] Document gotcha about training flag (#68915)

jubinchheda committed 2 years ago

27228656

dbr quant overhead [13/x]: cache results of get_module_hook_type (#68841)

vkuzo committed 2 years ago

f253370b

dbr quant: fix debugging fqn info for converted model (#68840)

vkuzo committed 2 years ago

2ad4727a

dbr quant overhead[12/x]: turn off overrides for module convert output hook (#68839)

vkuzo committed 2 years ago

a03fe9ba

dbr quant: remove unnecessary outputs hook (#68838)

vkuzo committed 2 years ago

515db567

dbr quant overhead[11/x]: speed up module convert hook (#68837)

vkuzo committed 2 years ago

e3af582f

dbr quant overhead[10/x]: disable torch_function overrides for leaf nodes (#68836)

vkuzo committed 2 years ago

be70477a

Add ModuleInfo-based device transfer tests (#68092)

thomasjpfan committed 2 years ago

1342f19a

Sparse CSR CUDA: Add torch.sparse.sampled_addmm (#68007)

IvanYashchuk committed 2 years ago

89a145fd

Newer Older