pytorch/xla

Pull Requests Commits

JackCaoG committed 1 year ago

b5702031

Support fullgraph for torch_xla.compile (#7776)

JackCaoG committed 1 year ago

Verified 731eb08b

[Pallas] Fix the doc (#7788)

alanwaketan committed 1 year ago

Verified 942ba7a6

Update libtpu_storage_path (#7786)

bhuvanpkaruturi committed 1 year ago

Verified ef49ddd8

Update GPU Development Doc for Hermetic Python setup (#7781)

lsy323 committed 1 year ago

Verified af7d2ea6

Further reduce number of compiling procs (#7779)

lsy323 committed 1 year ago

Verified e52b53f1

reenable dynamo dynamic shape test (#7775)

JackCaoG committed 1 year ago

Verified b09c9699

Update eager doc to use torch_xla.compile (#7778)

JackCaoG committed 1 year ago

Verified 28de4eb3

Disable open PJRT for xal2 if OS env disable flag is true (#7769)

FanhaiLu1 committed 1 year ago

Verified fb2d4e14

disable dynamo dynamic shape test since it is broken by some upstream… (#7773)

JackCaoG committed 1 year ago

Verified 622f4f93

openxla pin update (#7758)

zpcore committed 1 year ago

Verified dc3b265a

clean up deprecated (#7753)

zpcore committed 1 year ago

Verified 5702dae1

Refactor torch_xla.experimental.compile and move it out of experimental (#7750)

JackCaoG committed 1 year ago

Verified bdd00e54

Make libtpu install more explict (#7737)

JackCaoG committed 1 year ago

Verified 009e31aa

Fix few ops (#7759)

qihqi committed 1 year ago

Verified 2b5538d6

Add a guide to set up Visual Studio Code with PyTorch/XLA (#7755)

will-cromar committed 1 year ago

Verified b2e5c7d2

Move fsdp autowrap test to v4 TPUCI only (#7751)

JackCaoG committed 1 year ago

Verified db7fada8

Fix tril (#7744)

qihqi committed 1 year ago

Verified f7413b43

Remove `torch_xla[tpuvm]` optional dep (#7736)

will-cromar committed 1 year ago

Verified 44d637ad

[torch_xla2] Implement `bessel_*`, `chebyshev_*`, `hermite_*`, `erfcx`, `i0e` (#7723)

will-cromar committed 1 year ago

Verified 07a83636

add test for dynamo + traceable collectives (#7745)

JackCaoG committed 1 year ago

Verified b9c19b45

[benchmarks] Turn CUDAGraphs on for inductor. (#7749)

ysiraichi committed 1 year ago

Verified 0b990857

[benchmarks] Set matrix multiplication precision. (#7748)

ysiraichi committed 1 year ago

Verified 2870e93b

[benchmarks] Fix batch size logic. (#7747)

ysiraichi committed 1 year ago

Verified 62807012

Support `@xla.step` decorator without parentheses (#7746)

will-cromar committed 1 year ago

Verified 2e1ebfc0

fix TPU test regarding word_size (#7743)

zpcore committed 1 year ago

Verified d77bf6fe

Fix the crash when symint is part of the output with dynamic torch.compile (#7735)

JackCaoG committed 1 year ago

Verified df59b1c3

Use torch_xla.experimental.compile for all examples (#7642)

JackCaoG committed 1 year ago

Verified 735dd435

Fix the SPMD doc strng (#7718)

JackCaoG committed 1 year ago

Verified b4d82588

Update readme with 2.4 release version (#7742)

bhavya01 committed 1 year ago

Verified 2c770683

Older