strisunshinewentingwang/pytorch

Pull Requests Commits

Enabled cumsum and cumprod for bool tensors (#23346)

izdeby committed 5 years ago

cf0f3556

Open source 3D depthwise conv (#23164)

Du Tran committed 5 years ago

c9312e1a

Specifying libtorch variant in libtorch builds (#23348)

pjh5 committed 5 years ago

73dee44e

Switch keys to be sequential and stable in pickle serialization

zdevito committed 5 years ago

3297d8e2

Fix pickler bug where it would not load if no tensors were saved

zdevito committed 5 years ago

93da1030

Memoize storages in pickler

zdevito committed 5 years ago

7922b505

Unwrap DataParallel automatically (#23334)

houseroad committed 5 years ago

71a047c3

Skip QNNpack tests on ppc64le (where support is not enabled) (#23343)

dncliss committed 5 years ago

c23ba350

Improve the error message for ONNX export (#23317)

houseroad committed 5 years ago

22c169fb

max_pool_with_indices: return valid indices if all input elements are -inf (#23161)

skrah committed 5 years ago

87d3f665

add notes about overshoot in bicubic mode (#23321)

Ailing Zhang committed 5 years ago

b7d90332

BlackBoxPredictor OSS part N: ThreadLocalPtr, InferenceGraph (#23257)

salexspb committed 5 years ago

d522b3ca

Move OptionalType wrapping out of constants.cpp

davidriazati committed 5 years ago

2915d530

Better error for compiling a module type

davidriazati committed 5 years ago

48ca64db

Add docs about prod ecosystem features (#23010)

Dmytro Dzhulgakov committed 5 years ago

d6dcec37

Remove torch::autograd::Node::get_shared_ptr()

mal committed 5 years ago

87482bb1

Support LSTM with FP16 weight (#23291)

mingzhe09088 committed 5 years ago

8fdbe1e1

Revert D16440000: [pytorch][PR] Re-land "Fix error message for a wrong fork CUDA"

ezyang committed 5 years ago

1f608d09

Prevent xla-build from clobbering pytorch_linux_trusty_py3_6_gcc5_4_test

zou3519 committed 5 years ago

1a9edfcd

deepCopy also copies type information of lists (#23271)

smessmer committed 5 years ago

0c0ffccb

Revert D16441000: Switch from KaTeX to imgmath for documentation rendering.

ezyang committed 5 years ago

895e79ad

Quantized conv avoid functional usage (#22733)

Zafar Takhirov committed 5 years ago

94711d74

Reduce number of processes spawned for gloo_test.TestCase.test_forked_cw (#23221)

pietern committed 5 years ago

67179d71

Fix argument names in torch doc (#22973)

Will Feng committed 5 years ago

3ed79f4b

Revert D16423217: [pytorch][PR] Update sleef to master, fixes #20535

pjh5 committed 5 years ago

eb51131f

Fix the problem in parseOctal and throw exception if use \xhh to specify hex value (#23198)

houseroad committed 5 years ago

803af998

Suppress warnings in tests

davidriazati committed 5 years ago

b9a7fc52

Enabled 'add_cuda' for bool and fixed alpha scalar bug (#23044)

izdeby committed 5 years ago

200cb836

Change TensorIterator to be stack allocated, using named return value optimization to elide copies.

Owen Anderson committed 5 years ago

fbf28b54

Update sleef to master, fixes #20535 (#23168)

ezyang committed 5 years ago

7203612f

Newer Older