strisunshinewentingwang/pytorch

Pull Requests Commits

don't compile dnnlowp.cc in avx2 option (#15147)

jspark1105 committed 5 years ago

fab78827

docs: minor spelling tweaks

brettkoonce committed 5 years ago

d8260239

Export defs.bzl to open source for pytorch (#15132)

zdevito committed 5 years ago

2211a283

Add back c2 string_utils include header to benchmark_helper

bddppq committed 5 years ago

107c9ef5

use ROCm 1.9.2 fp16 capabilities in rocBLAS and MIOpen interfaces (#14994)

iotamudelta committed 5 years ago

6610ace2

Optimize CPU GenerateProposals op by lazily generating anchors (3-5x faster) (#15103)

viswanathgs committed 5 years ago

f34d8270

Implement torch.tril_indices and torch.triu_indices (#12653) (#14904)

mrshenli committed 5 years ago

90f9e810

Minor documentation mistake (#15068)

imran3180 committed 5 years ago

342e62f1

Add script standard library documentation + cleanup (#14912)

David Riazati committed 5 years ago

5837320b

Move adaptive avg pooling 2d to ATen native (#14714)

xnder committed 5 years ago

64b33642

Move numa.{h, cc} to c10/util (#15024)

jerryzh168 committed 5 years ago

63e77ab6

Stop erroneously running aten::warn (#15124)

zou3519 committed 5 years ago

b34ab435

Move CUDAGuard, CUDAStream and CUDAGuardImpl to c10/cuda (#14248)

ezyang committed 5 years ago

2d485ffb

Kill Type.storage. (#15075)

gchanan committed 5 years ago

9943cf23

fix infinite loop when get_max_threads is nonzero but num_threads is 1

Brennan Vincent committed 5 years ago

9d2955c3

Ensure there aren't variables in checked_tensor_unwrap, checked_tenso… (#15105)

gchanan committed 5 years ago

68ad9ae5

Add better support for bools in the graph fuser (#15057)

zou3519 committed 5 years ago

0ad39ec5

fix some tests that I accidentally disabled (#15077)

Brennan Vincent committed 5 years ago

f36a84b7

Don't setup x86_64-linux-gnu-gcc as an sccache wrapper. (#15078)

ezyang committed 5 years ago

3ae68426

Use c10::to_string that works cross platform (#15117)

bddppq committed 5 years ago

00a4c8d4

Add EmptyNameScope to allow you jump out from current scope. (#14631)

zhipingx committed 5 years ago

1423c0d9

Remove linker and dlopen flags that allowed undefined symbols in rocm build (#15091)

bddppq committed 5 years ago

479481b6

Fix serialization (#15033)

goldsborough committed 5 years ago

0dade986

Update the output format for benchmark_helper. It outputs the dimensi… (#15108)

sf-wind committed 5 years ago

e20f9bbe

Pre-commit flake8/clang-tidy (#15102)

zdevito committed 5 years ago

b07ee44f

add gloo support for gather on GPU (#14916)

janewangfb committed 5 years ago

f8455ed7

Fix include paths for UndefinedTensorImpl.h

smessmer committed 5 years ago

3fa53da6

Move UndefinedTensorImpl to c10 (meh) (#14817)

smessmer committed 5 years ago

63db95dd

Fix include paths for TensorImpl.h

smessmer committed 5 years ago

2dfdbef9

Move TensorImpl to c10 (yay!)

smessmer committed 5 years ago

9e9e87c1

Newer Older