pytorch/xla

Pull Requests Commits

benchmarks/aggregate: distinguish between openxla/_eval (#6159)

cota committed 2 years ago

Verified 0c1c68c3

Apply precision config env vars in the root process. (#6152)

golechwierowicz committed 2 years ago

Verified dfcf306e

Remove branch on undefined variable. (#6072)

ysiraichi committed 2 years ago

Verified c3e341c7

[benchmarks] Collect XLA compilation and execution metrics. (#6153)

ysiraichi committed 2 years ago

Verified 4103eb65

Mark uuid as constant in StableHLOCompositeBuilder (#6174)

chunnienc committed 2 years ago

Verified 2ac4cbc6

bump ci build timeout (#6169)

lsy323 committed 2 years ago

Verified 1e87f8f5

Fix precision issue of pow(int, float) (#6103)

qihqi committed 2 years ago

Verified e500129a

Don't explcitly include virtual device in hash (#6148)

jonb377 committed 2 years ago

Verified d2e06760

Reapply with fix "Add comments to _is_available and raise exceptions … (#6143)

frgossen committed 2 years ago

Verified f9dc8249

Extract fingerprinting for experiment and model config (#6125)

frgossen committed 2 years ago

Verified 59248c47

Add comments in benchmark_experiment and generalize workaround for too long strings in `filename_str` (#6117)

frgossen committed 2 years ago

Verified 29b183b3

Add more benchmark tests for openxla and openxla_eval backends (#6114)

frgossen committed 2 years ago

Verified 761b5d9e

Move `test_result_analyzer.py` to the test directory. (#6130)

ysiraichi committed 2 years ago

Verified 3f3068b1

Support forwarding dynamic shape to tf saved model (#6149)

qihqi committed 2 years ago

Verified bf5edb39

Fix zero-1 bug for inferring local ranks (#5936)

YangFei1990 committed 2 years ago

Verified 5577dd78

Add mark_tensor variant taking torch.Tensor id (#6132)

lsy323 committed 2 years ago

Verified d6d8d477

Revert "Initialize compilation env hash" (#6147)

jonb377 committed 2 years ago

Verified 6aaca152

Initialize compilation env hash (#6140)

jonb377 committed 2 years ago

Verified c4f8772d

Do manual type conversion for isnan op (#6110)

wonjoo-wj committed 2 years ago

Verified 5c12670a

Uncomment test_aten_isinf_2 at test_core_aten_ops.py (#6105)

wonjoo-wj committed 2 years ago

Verified fe7e1fde

Revert "Add comments to _is_available and raise exceptions in unimple… (#6141)

frgossen committed 2 years ago

Verified ddc2e34a

Add flag for equal_nan in test_core_aten_ops.py (#6104)

wonjoo-wj committed 2 years ago

Verified e326ed29

Do not fail on lack of default precision set. (#6139)

golechwierowicz committed 2 years ago

Verified 6b1344e6

ci: install pandas (#6137)

cota committed 2 years ago

Verified a2f80e46

benchmarks/aggregate: remove timestamp from "latest" title (#6138)

cota committed 2 years ago

Verified 9f9c9088

Add comments to _is_available and raise exceptions in unimplemented cases (#6118)

frgossen committed 2 years ago

Verified 491cbed7

Revert "Fix model initialization. (#6076)" (#6134)

frgossen committed 2 years ago

Verified 788e1b5e

Apply default precision from bench models to XLA (#6133)

golechwierowicz committed 2 years ago

Verified a0856820

Re-land experiment runner tests (#6113)

frgossen committed 2 years ago

Verified 3bbafc2a

Enforce successful imports and cause failure if not set up correctly (#6119)

frgossen committed 2 years ago

Verified ad02e793

Newer Older