strisunshinewentingwang/pytorch

Pull Requests Commits

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

fix compilation order for class methods (#20094)

suo committed 5 years ago

e37d9c81

Rearrange stopping condition in CompositeReader (#20062)

kittipatv committed 5 years ago

0aa7407d

Export randn_like in ONNX exporter (#20093)

houseroad committed 5 years ago

b3c35e52

Index into a tuple with non constant integer (#20081)

Elias Ellison committed 5 years ago

26f52756

Cleanup includes in torch/csrc/* (#19924)

Mikhail Zolotukhin committed 5 years ago

722eb48f

Cleanup includes in torch/csrc/autograd/* (#19923)

Mikhail Zolotukhin committed 5 years ago

6ca38d98

Cleanup includes in torch/csrc/jit/* (#19922)

Mikhail Zolotukhin committed 5 years ago

8b469383

Newer Older