pytorch/xla

Pull Requests Commits

Update dynamo doc

JackCaoG committed 2 years ago

953205ca

remove torchvision in stablehlo.py (#5501)

lsy323 committed 2 years ago

Verified 1c42ba74

Add explict error message when tensor is on CPU for dynamo backend (#5499)

JackCaoG committed 2 years ago

Verified 9557fb72

Add a API to explictly init runtime (#5500)

JackCaoG committed 2 years ago

Verified e1ee592d

Support tuples in partition spec (#5488)

jonb377 committed 2 years ago

Verified 7a63c276

Refactor stablehlo API and put them in official location. (#5493)

qihqi committed 2 years ago

Verified 17abce76

Skip a couple tests on TPU due to precision issue (#5496)

JackCaoG committed 2 years ago

Verified d3d6ca30

Always use ExecuteReplicated with SPMD (#5494)

jonb377 committed 2 years ago

Verified 03991d44

Support input sharding changed after first dynamo tracing (#5477)

JackCaoG committed 2 years ago

Verified d1212e90

[SPMD] Make IR sharding custom sharding op (#5433)

alanwaketan committed 2 years ago

Verified dea9b296

[original author: mrnikwaws] Neuron operator support (#5471)

aws-kingrj committed 2 years ago

Verified 3e370b50

[SPMD] Patch nn.Linear (#5491)

alanwaketan committed 2 years ago

Verified 4f378e69

Add topological sorting to dynamo partitions (#5472)

wonjoo-wj committed 2 years ago

Verified b58327e1

Add ShardingSpec to XLATensor when it is created with a PJRTShardedData (#5489)

JackCaoG committed 2 years ago

Verified 6270cbae

Update artifacts.auto.tfvars for 2.1 release (#5483)

wonjoo-wj committed 2 years ago

Verified 014caa53

[Dist] Refactor ZeRO-1 (#5145)

hgt312 committed 2 years ago

Verified 059f6b9d

[SPMD] Fix XLA_DUMP_POST_OPTIMIZATIONS test (#5485)

alanwaketan committed 2 years ago

Verified f1e1e271

[SPMD] Fix TPU CI after #5478 (#5487)

alanwaketan committed 2 years ago

Verified 8a98a887

Add Python API to execute StableHLO bytecode (#5476)

lsy323 committed 2 years ago

Verified a2fab36f

Add @yeounoh to .github CODEOWNERS (#5482)

yeounoh committed 2 years ago

Verified 2d0a729f

[SPMD] Properly skip tests on TPU V2 (#5479)

alanwaketan committed 2 years ago

Verified 364819a9

[SPMD] Fix HybridMesh ordering (#5478)

alanwaketan committed 2 years ago

Verified b8b48451

Don't move full tensor to device in deferred_init (#4819)

Zach Zheng committed 2 years ago

Verified c57370dc

Enable PJRT C API Client and other changes for Neuron (#5428)

aws-kingrj committed 2 years ago

Verified daed1d0a

Introduce `torch_xla.runtime.use_spmd()` (#5474)

yeounoh committed 2 years ago

Verified 462cb0f4

Enabling sharding device data IR (#5475)

JackCaoG committed 2 years ago

Verified 9a5fdf39

Fix the input sharding for dynamo (#5469)

JackCaoG committed 2 years ago

Verified bf997819

bump the timeout for CI (#5470)

JackCaoG committed 2 years ago

Verified 76495f0b

Fix data type in Pow with Scalar base and Tensor exponent (#5467)

lsy323 committed 2 years ago

Verified 909f28fa

update remote cache key in ansible (#5463)

lsy323 committed 2 years ago

Verified 028df4da

Older