strisunshinewentingwang/pytorch

Pull Requests Commits

[jit] Add type tags to lists/dicts in pickle (#33255)

davidriazati committed 4 years ago

99e211e6

Apply clang-format to RPC files (#34139)

mrshenli committed 4 years ago

7da24b36

Use double quotes in C++ to stay consistent with Python RPC docs (#34095)

mrshenli committed 4 years ago

3af0dffe

Improve ProcessGroup RpcBackendOptions Constructor API (#34081)

mrshenli committed 4 years ago

f1085a8e

[Aten] Suppress valgrind leaks in libcuda (#34169)

malfet committed 4 years ago

9d1c971b

Make DEBUG == REL_WITH_DEB_INFO on CUDA build (#34153)

zasdfgbnm committed 4 years ago

1beb309e

.circleci: Re-do run nightly pipelines on tag (#34148)

seemethere committed 4 years ago

cb3905e8

Remove deprecated codepath for old-style autograd.Function (#30696) (#33956)

lixinyu committed 4 years ago

7cda964e

[JIT] Add modulelist indexing for integer literal (#29236)

Elias Ellison committed 4 years ago

04378eb6

Turn on strict dtype checking for test_torch.py (#33825)

ezyang committed 4 years ago

ba1bd417

Revert D20171428: [profiler] fix chrome tracing for profiler run with cuda

rohan-varma committed 4 years ago

c5799766

Completely kill CUDA_tensor_apply3 (#34026)

zasdfgbnm committed 4 years ago

f299c2d6

Migrate lerp from CUDA_tensor_apply3 to TensorIterator (#34025)

zasdfgbnm committed 4 years ago

1affaf8d

Migrate bce loss from CUDA_tensor_apply3 to TensorIterator (#34023)

zasdfgbnm committed 4 years ago

27f56632

[gloo] dont hold locks in calls to buffer in ProcessGroupGloo:RecvWork::wait() and (#33926)

rohan-varma committed 4 years ago

92083f31

[profiler] fix chrome tracing for profiler run with cuda (#33987)

rohan-varma committed 4 years ago

c93b1d42

Remove use of `.data` from optimizers (#33640)

Enealor committed 4 years ago

6a97777f

[fix] flake8 lint error (#34146)

jerryzh168 committed 4 years ago

f26bbb5f

Fix HistogramObserver to not do detach on input (#34114)

Dmytro Dzhulgakov committed 4 years ago

a8fc3d8c

[caffe2] fix ambiguous call to 'fmaxType' THCHalfAutoNumerics.cuh (#33569)

Igor Sugak committed 4 years ago

9650253d

fix sph batchnorm to use sph fma

Hector Yuen committed 4 years ago

49586a2a

Minimum build should also exclude XNNPACK (#34110)

xuhdev committed 4 years ago

49921cad

randn and normal_ for complex tensors (#34037)

anjali411 committed 4 years ago

fbc9c61c

Add API for listing functions overridable by __torch_function__ (#33791)

ngoldbaum committed 4 years ago

ad2825a2

improved TorchScript traceback (#33834)

zdevito committed 4 years ago

358450e0

In torch_test, mark every test that takes >5s on a DEBUG CPU-only build as slow test (#33901)

ezyang committed 4 years ago

74a0663a

CUDA Vectorized Dropout (#33879)

slayton58 committed 4 years ago

9b527b35

[pytorch][mobile] make sure mobile build work with dynamic dispatch (#34038)

ljk53 committed 4 years ago

0cf34cf6

[pytorch][CI] end-to-end custom build script (#34012)

ljk53 committed 4 years ago

51936c5e

[quant][graphmode] Observing input/output values in call site (#33277)

jerryzh168 committed 4 years ago

5b9f1ada

Newer Older