Intel® Extension for PyTorch* v1.12.300-cpu Release Notes

EikanWang released this 14 Sep 23:20

· 11 commits to release/1.12 since this release

Highlights

Optimize BF16 MHA fusion to avoid transpose overhead to boost BERT-* BF16 performance #992
Remove 64bytes alignment constraint for FP32 and BF16 AddLayerNorm fusion #992
Fix INT8 RetinaNet accuracy issue #1032
Fix Cat.out issue that does not update the out tensor (#1053) #1074

Full Changelog: v1.12.100...v1.12.300

Assets 8