pytorch/xla

Pull Requests Commits

Changed the Half::from_bits PR to use constexpr function.

dlibenzi committed 6 years ago

194c0e93

Use XLA build versioning only when needed, to avoid spamming the Conda env folder.

dlibenzi committed 6 years ago

45855192

Add inplace version for ne, eq

vishwakftw committed 6 years ago

1909c002

Use at::rand for GetTestTensor, remove GetTestTensor

vishwakftw committed 6 years ago

5bf2fe7d

Remove redundant self_tensor instantiation

asuhan committed 6 years ago

e56ea1b1

Add more arithmetic with scalar to ATen XLA tensor

asuhan committed 6 years ago

d23cfaff

Fix formatting of aten_xla_type.cpp

asuhan committed 6 years ago

ef1accea

Merge pull request #385 from ailzhang/fix_concat_negative_index

ailzhang committed 6 years ago

Verified c8a7038a

Merge branch 'master' of https://github.com/pytorch/xla into fix_concat_negative_index

Ailing Zhang committed 6 years ago

15c2947b

address comments

Ailing Zhang committed 6 years ago

8108b69e

Allow setting/overriding the XLA tensor ScalarType.

dlibenzi committed 6 years ago

d8c558c2

Ailing Zhang committed 6 years ago

30b158b3

Make Cast IR op use the convert facility.

dlibenzi committed 6 years ago

4c67cfbe

Add tests for inplace ge, gt, le, lt

vishwakftw committed 6 years ago

9ef38116

Add inplace versions for ge, gt, lt, le

vishwakftw committed 6 years ago

686aa82a

Adjust to Undefined removal in core PyTorch

asuhan committed 6 years ago

d10e3f4d

Fix rand() infinite recursion with XLA device

asuhan committed 6 years ago

b09c611c

New C10 visibility fix ... once again.

dlibenzi committed 6 years ago

87c6fe24

Added aten::fmod on at::Scalar type.

dlibenzi committed 6 years ago

ad4416ff

debug torchvision

Ailing Zhang committed 6 years ago

ca01ee5f

Added atan2 operation

vishwakftw committed 6 years ago

77f500c6

Add inplace version for sqrt, rsqrt, reciprocal, erf, erfinv, erfc

vishwakftw committed 6 years ago

42a3db01

Move cross implementation to tensor_ops module

asuhan committed 6 years ago

886ac817

Add cross to ATen XLA tensor

asuhan committed 6 years ago

603b5b38

Remove patch which landed

asuhan committed 6 years ago

edcf112b

Pin torchvision to the last working version

asuhan committed 6 years ago

58d5740f

Add sort to ATen XLA tensor

asuhan committed 6 years ago

a4cb8054

Allow remaining binary arithmetic operations with a non-XLA tensor

asuhan committed 6 years ago

b58cf57e

Add inplace version for sign, ceil, floor, abs and neg

vishwakftw committed 6 years ago

eff84f56

Allow the override of C10 visibility.

dlibenzi committed 6 years ago

0ff02ca0

Older