pytorch/xla

Pull Requests Commits

Update tensor_methods.cpp

ManfeiBai committed 1 year ago

Verified 4c834392

Update tensor_methods.h

ManfeiBai committed 1 year ago

Verified 0b5d56f8

ManfeiBai committed 1 year ago

Verified 2d04c52a

ManfeiBai committed 1 year ago

Verified 90c637dd

Update aten_xla_type.cpp

ManfeiBai committed 1 year ago

Verified 3419560c

Update xla_native_functions.yaml

ManfeiBai committed 1 year ago

Verified 5da0fcd9

Add docs on how to integrate with dynamo (#6459)

changm committed 1 year ago

Verified d0dd1a7f

[fix-up] nightly.sh: define BM_DIR earlier in the file (#6487)

cota committed 1 year ago

Verified 14b1ee7d

nightly.sh: run llama2 benchmarks (#6485)

cota committed 1 year ago

Verified c9d38348

Get stablehlo version programmatically (#6476)

lsy323 committed 1 year ago

Verified 99b56c43

Remove assert on zero_point dtype of quantize/dequantize op (#6475)

lsy323 committed 1 year ago

Verified 81284808

benchmarks: add script to run llama2 inference benchmarks (#6481)

cota committed 1 year ago

Verified a9dc9f12

aggregate.py: add format=latest_grouped (#6472)

cota committed 1 year ago

Verified f37c4542

Infer output dtype for rsub using generic template (#6473)

lsy323 committed 1 year ago

Verified 9884e05a

Fix some more core aten ops (#6462)

wonjoo-wj committed 1 year ago

Verified e8dfc865

Add more ops to torch_xla2 (#6474)

qihqi committed 1 year ago

Verified 6a0cb712

Wrap constant arg in XLAExportInterpreter (#6460)

lsy323 committed 1 year ago

Verified 5dbbb284

bring TPU CI bakc to green by reinstall torch/torchvision (#6458)

ManfeiBai committed 1 year ago

Verified e6baedd5

Add process group documentation for SPMD (#6469)

jonb377 committed 1 year ago

Verified 732a1c7f

moving torch_xla2 experimental code in-tree (#6463)

qihqi committed 1 year ago

Verified 4814d64c

Support multiple StableHLO Composite outputs (#6295)

chunnienc committed 1 year ago

Verified 535d398b

Fix global_device_count(), local_device_count() for single process on CUDA (#6022)

vanbasten23 committed 1 year ago

Verified 8fc8d57a

Reenable TestPoolingNNDeviceTypeXLA (#6453)

qihqi committed 1 year ago

Verified 3e684095

Force guards when performing boolean operations on XLASymNode (#6433)

ezyang committed 1 year ago

Verified fe604e97

[benchmarks] Add support for AMP execution. (#6447)

ysiraichi committed 1 year ago

Verified ad822bf0

torchbench_model.py: fix YAML skip file path (#6457)

cota committed 1 year ago

Verified 510bb8ea

Unify TORCH_LIBRARY definitions (#6455)

alanwaketan committed 1 year ago

Verified 19923bd1

Disable test_EmbeddingBag_per_sample_weights_and_new_offsets (#6456)

JackCaoG committed 1 year ago

Verified 21653dc9

Fix benchmark process terminaion issue (#6441)

zpcore committed 1 year ago

Verified d1a454fe

Disable TestPoolingNNDeviceTypeXLA temporarily. (#6452)

yeounoh committed 1 year ago

Verified d0a5c34e

Older