pytorch/xla

Pull Requests Commits

Add test for nested to-XLA tensor conversions.

dlibenzi committed 6 years ago

5a232c34

Expand rank 0 indices for index_{add, copy, fill, select}

asuhan committed 6 years ago

994b6b38

Merge pull request #938 from ailzhang/fix_911

ailzhang committed 6 years ago

Verified b84f9a1b

rename variables

Ailing Zhang committed 6 years ago

0dfa7a7b

Update README.md

jysohn23 committed 6 years ago

9d794192

add python test case

Ailing Zhang committed 6 years ago

ef38277a

add override for empty_strided

Ailing Zhang committed 6 years ago

203e8a2b

Merge pull request #937 from ailzhang/remove_auto_test_tensor

ailzhang committed 6 years ago

Verified 83f71824

remove undesired auto usage in test_tensor

Ailing Zhang committed 6 years ago

73b6fbeb

Merge pull request #934 from ailzhang/fix_batchnorm

ailzhang committed 6 years ago

Verified 2f3c47cb

fix auto usage; fix test

Ailing Zhang committed 6 years ago

91269b0f

RateTracker smooth every time we report instead of every time we update (#930)

taylanbil committed 6 years ago

Verified 8a5ddb57

split to 2 lines

Ailing Zhang committed 6 years ago

add320a3

Re-install `torchvision` when updating wheels (#935)

jysohn23 committed 6 years ago

Verified 1b9e49f2

address comment

Ailing Zhang committed 6 years ago

e155469e

Fix Cum{Sum,Prod} casting tests to use AllClose() instead of AllEqual().

dlibenzi committed 6 years ago

8bb5946b

allow return undefined XLATensor; add test case for batchnorm

Ailing Zhang committed 6 years ago

139ea519

Do not expand scalars in nll-loss Iota computations.

dlibenzi committed 6 years ago

df026adc

Change the Replication barrier implementation.

dlibenzi committed 6 years ago

314d2d05

Implement server side cluster resolving (#926)

jysohn23 committed 6 years ago

Verified 9418cbc5

address comment

Ailing Zhang committed 6 years ago

b20e3345

Ailing Zhang committed 6 years ago

a361e16f

fix edge case batch=0

Ailing Zhang committed 6 years ago

b349b6a7

Ailing Zhang committed 6 years ago

d3158d58

Ailing Zhang committed 6 years ago

1d71bbbf

Wire in dense gather.

dlibenzi committed 6 years ago

b04b86ee

Make aten::as_strided() as a view.

dlibenzi committed 6 years ago

750c39b8

Enable multi TCP stream support in XRT.

dlibenzi committed 6 years ago

6428b39f

Bring the latest TF code with multi TCP stream support and new torch gather.

dlibenzi committed 6 years ago

d2a827c1

Implement client side cluster resolving (#910)

jysohn23 committed 6 years ago

Verified 0d0f3f8d

Older