pytorch/xla

Pull Requests Commits

Revert "Update cpp test"

JackCaoG committed 3 years ago

d2d03ad2

Update cpp test

JackCaoG committed 3 years ago

9b674c20

Implement the remaining backend interface (#3627)

JackCaoG committed 3 years ago

Verified e2235ff4

Use tcmalloc on TPUVM docker (#3647)

JackCaoG committed 3 years ago

Verified a5d83f41

Load libtpu in XRT local service (#3645)

will-cromar committed 3 years ago

Verified fe5f2db8

code for async scalar transfer (#3618)

aws-rhsoln committed 3 years ago

Verified 01e3c964

Cleanup README (#3634)

JackCaoG committed 3 years ago

Verified 4650c0ec

Fix ScopePusher regression issue (#3637)

wonjoo-wj committed 3 years ago

Verified f72dcc65

Install libtpu on TPUVM docker (#3635)

JackCaoG committed 3 years ago

Verified 54b4ddcf

fix torchvision pinning (#3643)

bdhirsh committed 3 years ago

Verified 1f9d52c7

Revert "Revert "Index device error (#3632)" (#3640)" (#3641)

bdhirsh committed 3 years ago

Verified 7ae5f50f

Revert "Index device error (#3632)" (#3640)

JackCaoG committed 3 years ago

Verified 7c99ad06

Index device error (#3632)

bdhirsh committed 3 years ago

Verified 9d5f6f68

delegate sym_sizes to sizes since Pytorch is now using sym_sizes (#3636)

Krovatkin committed 3 years ago

Verified 4f9d3bdc

replace out= lowerings with functional lowerings (#3620)

bdhirsh committed 3 years ago

Verified 9d8706f6

Build nightly torchaudio wheel (#3626)

JackCaoG committed 3 years ago

Verified 943f38f8

Refactor PjRt tests to take advantage of dynamic libtpu loading. (#3628)

will-cromar committed 3 years ago

Verified 3b2bbc8f

More conservative IsSparseGather check to favor dense gather on TPU (#3629)

yeounoh committed 3 years ago

Verified 91095239

Add TPUVM wheel

JackCaoG committed 3 years ago

0c8a82d6

Update nightly version (#3621)

JackCaoG committed 3 years ago

Verified 8ae2ec50

Delete .torch_pin

JackCaoG committed 3 years ago

Verified b821f4b5

Add torch.save(bf16) test (#3579)

JackCaoG committed 3 years ago

Verified 2544253e

Load libtpu when ComputationClient is created. (#3616)

will-cromar committed 3 years ago

Verified 8b6d4fa1

Revert "update tensor indexing to fail on `non_xla_tensor[xla_tensor]` (#3414)" (#3622)

JackCaoG committed 3 years ago

Verified ae61f2f6

Wrap ComputationClient::Computation inside of torch_xla::Computation (#3614)

JackCaoG committed 3 years ago

Verified 41a25f5d

update tensor indexing to fail on `non_xla_tensor[xla_tensor]` (#3414)

bdhirsh committed 3 years ago

Verified 83b5fe8d

XLA SPMD PoC implementation with XRT (#3476)

yeounoh committed 3 years ago

Verified 97ab3b87

Move more test under XLA_SKIP_MP_OP_TESTS flag (#3615)

JackCaoG committed 3 years ago

Verified 5dae54bc

Partially implement LTC BackendInterface (#3571)

JackCaoG committed 3 years ago

Verified 73aa5c5e

Update sparsity check heuristics used in TorchGather (#3610)

yeounoh committed 3 years ago

Verified e6e9f2e6

Newer Older