pytorch/xla

Pull Requests Commits

Fix the Compile A Palooza issue that arise after the PT Variable code change.

dlibenzi committed 6 years ago

9f6887ec

Override ATEN to() operations.

dlibenzi committed 6 years ago

0768145b

Fix aten_xla_type.h formatting

asuhan committed 6 years ago

ca378226

Update README.md

jysohn23 committed 6 years ago

7c3091cc

Clarify the required bazel version

apaszke committed 6 years ago

dbb11d00

Merge pull request #766 from ailzhang/fix_pt20994

ailzhang committed 6 years ago

Verified 49a20afe

fix breakage PT#20994

Ailing Zhang committed 6 years ago

12c3a7b3

Merge pull request #764 from pytorch/kaboom_rebased

ailzhang committed 6 years ago

Verified f4bd9cc2

removed unused import

taylanbil committed 6 years ago

47c50622

Change the code generation to follow new PyTorch tensor dispatch.

dlibenzi committed 6 years ago

ce6fad49

set correct tensoroption for to

Ailing Zhang committed 6 years ago

68c3a45a

Emit graph IDs in diffs.

dlibenzi committed 6 years ago

62cdf9ee

Handle resize in place of views.

dlibenzi committed 6 years ago

3cdaaabb

fix change introduced by PT#21852

Ailing Zhang committed 6 years ago

c368f8c2

adding deepcopy back in as .to modifies inplace

taylanbil committed 6 years ago

fa5ec829

Davide s changes instead of deepcopy

taylanbil committed 6 years ago

3588cf4d

deepcopy the network so the parallel versions initialize to the same set

taylanbil committed 6 years ago

9d2a36f8

Expand S64 check as the current one does not work (S64 is indeed generally supported on TPU - just not for reduce-window).

dlibenzi committed 6 years ago

a6f5b7af

Add graph count within diffs.

dlibenzi committed 6 years ago

e3e2afbe

Mark the roots within TEXT and DOT graph export (HLO already has them by default).

dlibenzi committed 6 years ago

5e67b786

Merge pull request #756 from pytorch/asuhan/fix_build

ailzhang committed 6 years ago

Verified c6900496

Add namespace to QScheme in generator script

asuhan committed 6 years ago

def86c05

Implement resize in place.

dlibenzi committed 6 years ago

6284512e

Remove the 4D cap on batch norm.

dlibenzi committed 6 years ago

b598545a

Fix caffe library names in cmake machinery.

dlibenzi committed 6 years ago

ea6d05dd

Install torchvision from source (#751)

jysohn23 committed 6 years ago

Verified e40ece2c

Fix build broken by the removal of generator() method

asuhan committed 6 years ago

12c2bd93

Allowing squeezenet after verifying it runs e2e

taylanbil committed 6 years ago

e2977658

Merge pull request #748 from ailzhang/remove_unwrap

ailzhang committed 6 years ago

Verified 56a49fac

Send cum{sum, prod} to CPU for long type

asuhan committed 6 years ago

9ca50a3a

Older