pytorch/xla

Pull Requests Commits

Add _argmin and _argmax to ATen XLA tensor

asuhan committed 6 years ago

3f62f05d

Fix lack of negative dimension wrap around for all operators

asuhan committed 6 years ago

cdc88836

Remove trivial XLATensor::GetCanonicalDimension

asuhan committed 6 years ago

9c3c260f

Bump relative and absolute tolerance for TestSymEig

asuhan committed 6 years ago

cf2d9754

Fix tests which uses the default XLA device directly.

dlibenzi committed 6 years ago

304e6442

Added aten::symeig operation

vishwakftw committed 6 years ago

c9632f13

Add selu, selu_, celu, celu_ to ATen XLA type

asuhan committed 6 years ago

4b213de6

Add elu and elu_ to ATen XLA tensor

asuhan committed 6 years ago

14ae2bf7

Fix tensor copy across devices operations.

dlibenzi committed 6 years ago

8251e011

Add remaining pow overloads to ATen XLA tensor

asuhan committed 6 years ago

5bc95144

Fix formatting of init_python_bindings.cpp

asuhan committed 6 years ago

cd4bdca9

Make GatherParameters more robust

asuhan committed 6 years ago

e6fe8119

Trim unused imports.

dlibenzi committed 6 years ago

c61a73b9

Added a tensor based parallel loader.

dlibenzi committed 6 years ago

51967bf7

Fix XLATensorImpl size w/r/t in-place shape changing operations

asuhan committed 6 years ago

14c9d2b6

Adjust to module changes in core PyTorch

asuhan committed 6 years ago

0cf0e37a

Make permutation and transpose be views

asuhan committed 6 years ago

d7414cd8

Added aten::qr operation

vishwakftw committed 6 years ago

df776231

Handle view updates which are simple reshapes.

dlibenzi committed 6 years ago

bebb28c0

Make Python XLA module take input functions to prepare for at::Tensor parallel loader code.

dlibenzi committed 6 years ago

83f1fed0

Fix the fixup script to make it work with Kokoro VMs.

dlibenzi committed 6 years ago

de986843

Handle permute in views.

dlibenzi committed 6 years ago

2ca55596

Move transpose implementation to ops

asuhan committed 6 years ago

4861b81c

Replace t with transpose of 0 and 1

asuhan committed 6 years ago

6bd34d35

Add transpose to ATen XLA tensor

asuhan committed 6 years ago

86557ae3

Add unsqueeze_ to ATen XLA tensor

asuhan committed 6 years ago

c093be8c

Add tril_, triu_ to ATen XLA tensor

asuhan committed 6 years ago

8f98fd06

Add in-place transpose to ATen XLA tensor

asuhan committed 6 years ago

f0ebf473

Remove one-use, one liner, tensor API.

dlibenzi committed 6 years ago

4e8c456a

Added aten::svd operation.

dlibenzi committed 6 years ago

b06ebd48

Newer Older