pytorch/xla

Pull Requests Commits

Merge branch 'master' into ManfeiBai-patch-76

ManfeiBai committed 1 year ago

Verified bbc7307b

[SPMD] Check the real physical device type if XlaDeviceType::SPMD (#6652)

yeounoh committed 1 year ago

Verified 46e22300

[OpenXLA|TSL] Replace `tsl/platform/stringpiece.h` with `absl/strings/string_view.h` (#6657)

ManfeiBai committed 1 year ago

Verified 9c5c6d91

Revert "Add missing toml package to post-submit CI" (#6682)

vanbasten23 committed 1 year ago

Verified 1dc78948

Add missing shell options to `test/tpu/run_tests.sh` (#6681)

will-cromar committed 1 year ago

Verified 3dbcbaab

Add Dockerfile for TPU CI self-hosted runner (#6678)

mbzomowski committed 1 year ago

Verified 6bcf4fd3

nightly.sh: do not include one-off jobs in generated reports (#6679)

cota committed 1 year ago

Verified 6fe19d3c

[Pallas] Support Flash Attention (#6658)

alanwaketan committed 1 year ago

Verified 9d4dcae7

Remove `kubectl` install from TPU CI script (#6672)

will-cromar committed 1 year ago

Verified 57b07c3e

ManfeiBai committed 1 year ago

Verified 866b489e

[Pallas] Support Dynamo (#6477)

alanwaketan committed 1 year ago

Verified ce8ee38e

ManfeiBai committed 1 year ago

Verified 839c8a49

Make `as_strided_copy` materialize a new tensor with `index`. (#6624)

ysiraichi committed 1 year ago

Verified 3abc21df

ManfeiBai committed 1 year ago

1576ac13

Update convolution_helper.h

ManfeiBai committed 1 year ago

93cc844f

Generate storage type with signless semantics (#6613)

sdasgup3 committed 1 year ago

Verified b8864fc5

Install missing toml package for post-submit CI failure. (#6670)

vanbasten23 committed 1 year ago

Verified ca2fde79

add exp for jax_jit (#6651)

qihqi committed 1 year ago

Verified 7f18db7d

Implement mark_sharding as a custom dynamo op (#6524)

wonjoo-wj committed 1 year ago

Verified a80f1d79

Add missing toml package to post-submit CI (#6664)

vanbasten23 committed 1 year ago

Verified 3ba5987a

Make XLAShardedTensor use default __torch_function__ (#6625)

JackCaoG committed 1 year ago

Verified 5a113aff

disable dynamic shape export test (#6654)

lsy323 committed 1 year ago

Verified 508aa260

Fix output tensor shape for argmin and argmax where keepdim=True and dim=None (#6536)

mrnikwaws committed 1 year ago

Verified a1ab7fd8

[benchmarks] Small fixes for benchmarking script. (#6632)

ysiraichi committed 1 year ago

Verified 4cde625c

Support building StableHLO composite with different attr value types (#6615)

chunnienc committed 1 year ago

Verified e385c2fd

handle non-tensor input in tf saved_model input placeholder (#6640)

lsy323 committed 1 year ago

Verified a26d3c07

Dynamo bridge should only sync input XLATensor during execution (#6636)

JackCaoG committed 1 year ago

Verified 169633f6

Skip pytorch test test_parallel_cow_materialize_error (#6637)

wonjoo-wj committed 1 year ago

Verified 49dbb8d3

benchmarks/llama.py: fix handling of symlinks in workspace_dir (#6630)

cota committed 1 year ago

Verified 215b8b87

update dynamo debug message (#6629)

JackCaoG committed 1 year ago

Verified 8be7f269

Older