pytorch/xla

Pull Requests Commits

Make --tpu flag only accept single TPU (#979)

jysohn23 committed 6 years ago

Verified 719fa926

add test_set_batch_size flag (#982)

zcain117 committed 6 years ago

Verified f1d11d56

Fix cross shape for zero-sized dimensions (#980)

asuhan committed 6 years ago

27b043ce

fix edge case in slice when start > end (#977)

ailzhang committed 6 years ago

Verified 8c8e6547

run scatter add on numeric operands (#976)

ailzhang committed 6 years ago

5067d0e3

send norm p=0 back to cpu (#975)

ailzhang committed 6 years ago

27029d8c

fix *_like constructors (#974)

ailzhang committed 6 years ago

9e0da660

Implement distributed executor. (#971)

jysohn23 committed 6 years ago

Verified 6e60add5

Fix depthwise conv (#973)

ailzhang committed 6 years ago

72bfe5d2

Remove unused local.

dlibenzi committed 6 years ago

e797b41c

Lower default rate tracker smoothing factor.

dlibenzi committed 6 years ago

ad6f0cdc

Simplify combiner functor to not take the builder (the XlaOps carry one if needed).

dlibenzi committed 6 years ago

5a84d061

Ensure no partial replication batches are issued.

dlibenzi committed 6 years ago

5a9e8d4a

make frac match pytorch python result (#967)

ailzhang committed 6 years ago

Verified cd2a1489

Merge pull request #962 from ailzhang/fix_922

ailzhang committed 6 years ago

Verified debee07d

address comment

Ailing Zhang committed 6 years ago

dbef9900

Ailing Zhang committed 6 years ago

fd86596a

Merge pull request #965 from pytorch/aifix_warnings

ailzhang committed 6 years ago

Verified cdd0a752

Fix build warnings.

dlibenzi committed 6 years ago

ef245db4

disable MKLDNN build in CI

Ailing Zhang committed 6 years ago

fac438da

Merge branch 'master' of https://github.com/pytorch/xla into fix_922

Ailing Zhang committed 6 years ago

e2d49bfb

Merge pull request #955 from ailzhang/lower_convolution

ailzhang committed 6 years ago

Verified b986a467

address comment

Ailing Zhang committed 6 years ago

dd29b79f

another shape that mkldnn doesn't like

Ailing Zhang committed 6 years ago

65e3d7d2

Unbreak pods initialization. (#963)

jysohn23 committed 6 years ago

Verified 698bc864

Merge branch 'master' of https://github.com/pytorch/xla into fix_922

Ailing Zhang committed 6 years ago

d6ae54ce

lower rrelu_with_noise_backward

Ailing Zhang committed 6 years ago

71c8bccf

Add Dimname, DimnameList to codegen to support named tensor build

zou3519 committed 6 years ago

4b8d8c0a

Ailing Zhang committed 6 years ago

947bcbc1

lower rrelu_with_noise

Ailing Zhang committed 6 years ago

158a9a4a

Newer Older