pytorch/xla

Pull Requests Commits

JackCaoG committed 2 years ago

292c28a9

Update xla_graph_executor.cpp

JackCaoG committed 2 years ago

Verified cb942993

Merge branch 'master' into JackCaoG/cache_sharded_placeholder_dynamo

JackCaoG committed 2 years ago

Verified a899ca5e

Remove base image override from TPU CI build (#5453)

will-cromar committed 2 years ago

Verified 63febd9e

parallelize SPMD inputhandler and GetDataShards (#5447)

JackCaoG committed 2 years ago

Verified 481d0b04

Revert "Install gcc-10 (#5450)" (#5452)

will-cromar committed 2 years ago

Verified 493bb734

Install gcc-10 (#5450)

qihqi committed 2 years ago

Verified 65b7639c

add missing wait deivce ops

JackCaoG committed 2 years ago

87d4dde0

Revert "Update gcc to 10. (#5445)" (#5449)

lsy323 committed 2 years ago

Verified 55249edd

remove dead code

JackCaoG committed 2 years ago

96018064

Update the random seed for every dynamo execution (#5444)

JackCaoG committed 2 years ago

Verified 13f9382c

JackCaoG committed 2 years ago

afc349c6

Update gcc to 10. (#5445)

qihqi committed 2 years ago

Verified 454e916e

address review comments

JackCaoG committed 2 years ago

1e344fb5

Cache the output sharding spec for dynamo

JackCaoG committed 2 years ago

aab2b80f

Move the C++ test build to CI build job instead of test job (#5442)

JackCaoG committed 2 years ago

Verified 4ef4dd34

Bugfix + add more test for llama (#5439)

qihqi committed 2 years ago

Verified a2a4ff4c

[PJRT] Support `torchrun` with `pjrt://` `init_method` (#5438)

will-cromar committed 2 years ago

Verified 9db8fac4

Update fix_includes.sh (#5441)

aazzolini committed 2 years ago

Verified 5b88b5f8

Remove TPU loading hack (#5435)

will-cromar committed 2 years ago

Verified 709c75a0

Add repo for libcudnn8=8.7.0.84 and CUDA 11.8 (#5425)

mateuszlewko committed 2 years ago

Verified f06c31f5

[PJRT|TPU] Update `test_xla_devices_single_process_all_chips` for expected device number (#5421)

ManfeiBai committed 2 years ago

Verified 2d97d3a1

[SPMD] named partition spec support (#5415)

yeounoh committed 2 years ago

Verified 85453d27

Fix the incorect clone arg condition in dynamo bridge (#5414)

JackCaoG committed 2 years ago

Verified 1835771c

SPMD cross slice-replication using partial_replication sharding (#5411)

yeounoh committed 2 years ago

Verified 6852df1d

Revert "Partially replicate lower-rank tensors (#5409)" (#5412)

yeounoh committed 2 years ago

Verified 628ba698

Partially replicate lower-rank tensors (#5409)

jonb377 committed 2 years ago

Verified 56a6a02a

Fix TPU collective ops test for multi-host TPUs (#5408)

will-cromar committed 2 years ago

Verified 2658c882

Refactor to share code between export_torch_model and save_as_stablehlo (#5388)

qihqi committed 2 years ago

Verified 2971efca

[SPMD] Multi-host batch sharded data loading (#5331)

khatwanimohit committed 2 years ago

Verified f2f877f8

Older