strisunshinewentingwang/pytorch

Pull Requests Commits

Merge branch 'master' of github.com:pytorch/pytorch into driazati/breakpad

David Riazati committed 3 years ago

e23fb399

Flip clip_grad_norm default for error_if_nonfinite to false (#55169)

maximsch2 committed 3 years ago

38a08a49

[JIT] Add recursive scripting for class type module attributes (#55124)

Meghan Lele committed 3 years ago

6866c033

Add interpolation kwarg to torch.quantile (#49267)

heitorschueroff committed 3 years ago

6e2d0200

skip test when appropriate

David Riazati committed 3 years ago

8b15c47a

[Gradient Compression] Update a warning in ddp_comm_hooks.rst (#55031)

Yi Wang committed 3 years ago

e5930447

[StaticRuntime] Unbreak benchmarks. (#55199)

Mikhail Zolotukhin committed 3 years ago

7ab53eb9

[PyTorch] Don't bother with SmallVector in TensorMaker (#55125)

swolchok committed 3 years ago

a0bb0968

Enhance Pipe docs to explicitly mention RPC initialization. (#55187)

pritamdamania committed 3 years ago

02af4b51

Replace AutoNonVariableTypeMode with InferenceMode in fbcode. (#55114)

Ailing Zhang committed 3 years ago

24c90495

Split copy_ kernel to InplaceOrView. (#55133)

Ailing Zhang committed 3 years ago

181de406

Don't globally disable any ShellCheck warnings (#55165)

samestep committed 3 years ago

09670c7d

Revert D25399470: add channels last for MaxPool2d

agolynski committed 3 years ago

978fca64

Modified lstsq_helper to accept lapack error codes tensor (#54720)

IvanYashchuk committed 3 years ago

e406d4e6

ns for fx: weight extaction for conv1d and conv3d (#55079)

vkuzo committed 3 years ago

8062545c

ns for fx: ensure kwargs are handled when graph matching (#55078)

vkuzo committed 3 years ago

80b1b7e4

ns for fx: clean up debug print statements (#55077)

vkuzo committed 3 years ago

a590fa7a

ns for fx: move it to top level file (#55060)

vkuzo committed 3 years ago

f6b25e75

ns for fx: weight extraction for nni.ConvReLU2d (#54335)

vkuzo committed 3 years ago

c6cb99a6

ns for fx: make input logging work for multi node subgraphs (#54327)

vkuzo committed 3 years ago

5319d17b

ns for fx: make input logging work for multi-node subgraphs (#54326)

vkuzo committed 3 years ago

b8019cee

ns for fx: refactor test cases (#54280)

vkuzo committed 3 years ago

cbcde790

ns for fx: add support for shadowing linear fp16 patterns (#54275)

vkuzo committed 3 years ago

757e3cbf

add channels last for MaxPool2d (#48917)

mingfeima committed 3 years ago

f43eb59a

Revert D27478436: Use tensorpipe::Buffer::device() instead of tensorpipe::Buffer::deviceType().

Mike Ruberry committed 3 years ago

bdb225e9

Use tensorpipe::Buffer::device() instead of tensorpipe::Buffer::deviceType().

Lucas Hosseini committed 3 years ago

3e185253

[ATen][qembeddingbag] Avoid tensor refcount bumps (#55023)

Hao Lu committed 3 years ago

61914cb2

[c10] Add default constructor to Maybeowned (#55128)

Hao Lu committed 3 years ago

93d0f636

Adds torch.* API section for TorchScript Lang Ref (#53236)

gmagogsfm committed 3 years ago

ec609e74

[PyTorch Edge] Provide a method ObservedOperators::getUnobservedOperatorList() so that model tracer can empty it out during tracing (#55017)

dhruvbird committed 3 years ago

271879fe

Older