pytorch/xla

Pull Requests Commits

stgpetrovic committed 2 years ago

Verified 3a90f5af

Disable failing `test_operations.py` tests on TPU (#4622)

will-cromar committed 2 years ago

Verified 83f425a1

Skip execution for extract_compiled_graph (#4612)

JackCaoG committed 2 years ago

Verified b7707fa1

Fix patch command in TPU CI (#4623)

will-cromar committed 2 years ago

Verified 574cfdab

Update TF pin to 2/13 (#4615)

will-cromar committed 2 years ago

Verified dbe596c7

Fix HLO dumping (#4619)

JackCaoG committed 2 years ago

Verified f70afa45

Remove reference to `torch._six` compat layer (#4613)

XuehaiPan committed 2 years ago

Verified d29eb67c

[FSDP] Use state_dict hook to save shard metadata for consolidating checkpoint (#4572)

YangFei1990 committed 2 years ago

Verified d13090c4

Support multihost SPMD execution (#4573)

jonb377 committed 2 years ago

Verified ced64569

[PJRT] Use TFRT TPU client by default (#4600)

will-cromar committed 2 years ago

Verified 16b4f03b

Add int as supported index tensor type (#4602)

ymwangg committed 2 years ago

Verified 249b8e69

Support multiple GPUs on a single host with PjRt (#4297)

jonb377 committed 2 years ago

Verified 1e011c60

Add the counter check for dynamo tests (#4603)

JackCaoG committed 2 years ago

Verified de2160a9

Fix fp64 div issue (#4595)

ymwangg committed 2 years ago

Verified 9232b60d

Revert "Rename random_.from to random_.from_int (#4578)" (#4605)

JackCaoG committed 2 years ago

Verified 2a576e55

Fix typo in SPMD imagenet script (#4601)

jonb377 committed 2 years ago

Verified 265c3a6d

[PJRT] Use PJRT by default if XRT is not configured (#4599)

will-cromar committed 2 years ago

Verified 43ee2eb9

Increase cloud docker build "push" timeout (#4597)

yeounoh committed 2 years ago

Verified 8949dd12

Add simple Transformer fine-tuning test (#4593)

will-cromar committed 2 years ago

Verified b401b2a9

Update the list of supported CUDA images. (#4591)

yeounoh committed 2 years ago

Verified 7006b60b

Revert tf's protobuf version update since it regress XRT (#4594)

JackCaoG committed 2 years ago

Verified f92ec76d

Rename random_.from to random_.from_int (#4578)

wonjoo-wj committed 2 years ago

Verified 0136b7ce

Add expand symint python test and re-enable some previously disabled dynamic shape tests. (#4581)

vanbasten23 committed 2 years ago

Verified 76b34fb9

Make it an error when ask for a non symbolic size of a symbolic tensor (#4553)

vanbasten23 committed 2 years ago

Verified f808e698

[PJRT] Slice literals to return correct dynamic shape (#4583)

will-cromar committed 2 years ago

Verified 4b740f2c

Update cuda 11.7 (#4587)

yeounoh committed 2 years ago

Verified a497db46

Revert "Update CUDA ver to 11.7.0 (#4584)" (#4586)

yeounoh committed 2 years ago

Verified b13618a1

Update CUDA ver to 11.7.0 (#4584)

yeounoh committed 2 years ago

Verified b9d91033

Actual build sccache from source (#4570)

huydhn committed 2 years ago

Verified 33b36a71

Skip git pull if we're in the xla repo (#4560)

steventk-g committed 2 years ago

Verified 9f0326a4

Newer Older