strisunshinewentingwang/pytorch

Pull Requests Commits

Ignore `nn::Functional` submodules in `nn::Module` serialization (#19740)

Will Feng committed 5 years ago

5099db08

Refactor ProcessGroupNCCL collective primitives (#18820)

mxw committed 5 years ago

61d48aa9

Install TensorBoard for doc generation (#19810)

orionr committed 5 years ago

e1ebf330

update Anaconda download link (#19794)

huangyanhua committed 5 years ago

bacc8815

Removing unused arg f from _model_to_graph(). (#19647)

Spandan Tiwari committed 5 years ago

dafee117

Multiple module outputs and multiple calls to backward (#19799)

pietern committed 5 years ago

0d8a3610

Allow passing lists as trace inputs.

efaust committed 5 years ago

dcfb5620

C++ changes toward libtorch and libcaffe2 unification (#19554)

kostmo committed 5 years ago

8f0603b1

More topi support (#19728)

Yinghai Lu committed 5 years ago

9d180e60

Update foxi version (#19793)

zrphercule committed 5 years ago

c182824f

Automatic update of fbcode/onnx to 22662bfd4dcc6baebf29e3b823a051676f991001 (#19790)

houseroad committed 5 years ago

20c22bca

Add devtoolset 8 (gcc 8) + glibc 2.26 + centos 7.5 rocm docker image (#19767)

bddppq committed 5 years ago

f0d493d2

TensorBoard support within PyTorch (#16196)

lanpa committed 5 years ago

98e312cf

Always enable autodiff check (#19787)

Ailing Zhang committed 5 years ago

97e80ab6

Automatic update of fbcode/foxi to 8f74bc4df3a4cfc69b1a3eadf62aa29d9961c72d AND update Glow AND update C2 (#19792)

jackm321 committed 5 years ago

48d5ab54

Profiler: add Self CPU Time Total, CPU time total and other general improvements (#19378)

salexspb committed 5 years ago

7a8bc85f

Quantized SumRelu (#19319)

Zafar Takhirov committed 5 years ago

6e06154c

Use the QTensor with QReLU (#19312)

Zafar Takhirov committed 5 years ago

76307667

Changing the rounding in the QTensor (#19714)

Zafar Takhirov committed 5 years ago

db900849

Fix env vars needed for devtoolset7 binaries

pjh5 committed 5 years ago

e814c110

Fixing update_s3_htmls for binaries

pjh5 committed 5 years ago

c5cca653

Fix case for `activations` attribute in nn.RNN ONNX export. (#19368)

Spandan Tiwari committed 5 years ago

9ef8eb4c

Remove duplicate inlineCallToCode (#19724)

zdevito committed 5 years ago

a425e1cb

Serialize first-class version of functions (#19723)

zdevito committed 5 years ago

330990d8

Trace directly into first-class module form. (#19722)

zdevito committed 5 years ago

6cb1b994

@torch.jit.script(fn) now is a torch.jit.Function (#19721)

zdevito committed 5 years ago

31524bda

pybind CompilationUnit and Function directly (#19720)

zdevito committed 5 years ago

12f7c2de

caffe2 | Use _aligned_free in WorkerPool destruction (#19751)

Oleg Bogdanov committed 5 years ago

bf5a5c2a

Bug fix in bound shape inferencer (#19729)

Yinghai Lu committed 5 years ago

65496e4e

Fall back to asking nvcc for detecting cuda version if no *cudaart* is found (#19741)

t-vi committed 5 years ago

556c8a30

Newer Older