strisunshinewentingwang/pytorch

Pull Requests Commits

Support operator overloading for UDT (#20033)

Elias Ellison committed 5 years ago

6606ac5d

Remove brew libomp from binary mac machines (#20228)

pjh5 committed 5 years ago

9c62280e

Fix in benchmark_test_generator (#20237)

Ilia Cherniavskii committed 5 years ago

eecf52b4

Remove flake8 E303 (too many blank lines) (#20225)

suo committed 5 years ago

8a375189

Fix test_namedtuple_return (#20212)

zasdfgbnm committed 5 years ago

ba3e6de4

Use ignore=dirty in submodules. (#20135)

zdevito committed 5 years ago

785583a4

PyTorch Profiler Shape aggregation support (#20035)

salexspb committed 5 years ago

864cfbc2

support onnx export rnn with batch_first=True (#19766)

BowenBao committed 5 years ago

831bd1c2

Make graph->param_node()->next() the first node (#19788)

zdevito committed 5 years ago

e58817fe

Enable ROCm multi-gpu with Gloo

bddppq committed 5 years ago

bc539845

Add proper __repr__ to LogSoftMax (#20018)

sdg committed 5 years ago

cf55670b

Working on component-wise transformations that mimic `torch.cat` and `torch.stack` (#11868)

ahmadsalim committed 5 years ago

b8256280

add process_group in convert_sync_batchnorm (#19240)

zhangliliang committed 5 years ago

f7a78688

Add DirichletFullyConnectedActor to Soft Actor-Critic

econti committed 5 years ago

2356fac9

Add Custom graph fusion (#18588)

bwasti committed 5 years ago

4ca325df

Intra-op parallel microbenchmarks for PT (#19997)

Ilia Cherniavskii committed 5 years ago

19e68865

Allow a non-OpenMP based build (#19749)

Ilia Cherniavskii committed 5 years ago

481b6d02

Initialize Caffe2 only when running Caffe2 benchmarks (#19980)

Ilia Cherniavskii committed 5 years ago

8c97f0b1

Support for non-contiguous tensors and arbitrary dtypes in PT benchmarks (#19993)

Ilia Cherniavskii committed 5 years ago

0c7e98b7

Add USE_CUDA macro in THD DataChannelGloo for non-GPU use case (#20186)

xw285cornell committed 5 years ago

839343a4

Add torch.nn.quantized.functional namespace (#20042)

jerryzh168 committed 5 years ago

8ca10d35

Update logic for folding onnx::Constant nodes. (#20109)

Spandan Tiwari committed 5 years ago

838ada3a

Fix NameError with PYTORCH_JIT=0 (#20120)

davidriazati committed 5 years ago

877b7c1b

temporarily disable devtoolset7 nightlies (#20174)

kostmo committed 5 years ago

1cfe15ef

Cleanup includes in c10/core/CPUAllocator.cpp. (#19885)

Mikhail Zolotukhin committed 5 years ago

4211f674

lower batchmm to non-diff optimization (#19987)

wanchaol committed 5 years ago

8fbde946

Add logging import and failing MLP (#20115)

orionr committed 5 years ago

0c5dc965

Add options to Operator to enable registration of alias analysis passes (#19382)

bwasti committed 5 years ago

035966d5

Specialize Optional[T] to T (or subtype for Tensor) or None when executing graph (#18407)

t-vi committed 5 years ago

5c9ab6f4

allow classes to be used in their own methods (#20106)

suo committed 5 years ago

47f5be16

Newer Older