pytorch/xla

Pull Requests Commits

ManfeiBai committed 1 year ago

Verified 522720dc

ManfeiBai committed 1 year ago

Verified 9a856394

Update tf_logging.h

ManfeiBai committed 1 year ago

Verified 18956a9f

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

ysiraichi committed 1 year ago

Verified 3abc21df

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

Update profiler.cc

ManfeiBai committed 1 year ago

Verified 2a4558c0

Replace tsl::Status to absl::Status in profiler.h

ManfeiBai committed 1 year ago

Verified 09bf3dfb

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

Add test for Gemma model (#6631)

qihqi committed 1 year ago

Verified 74ba968e

AliasWithBufferDonorContext wtih default value (#6628)

JackCaoG committed 1 year ago

Verified c99ed84c

Add UInt16/32/64 support (#6626)

lsy323 committed 1 year ago

Verified 0a2e5b5a

Use BERT for the HF Transformers example (#6584)

will-cromar committed 1 year ago

Verified 3c8c54e4

BufferDonor based aliasing enhancement (#6619)

JackCaoG committed 1 year ago

Verified 472d90d0

Change to correct directory before running v2 TPU CI (#6622)

will-cromar committed 1 year ago

Verified db64a454

Automatically detect CUDA if it isn't set and default to upstream torch.cuda number of devices (#6605)

changm committed 1 year ago

Verified cb4983e9

[benchmarks] Run some models with smaller batch sizes. (#6542)

ysiraichi committed 1 year ago

Verified cd47390f

Add API to donate input buffer for dynamo execution (#6587)

JackCaoG committed 1 year ago

Verified 3e2a23ca

Newer Older