add support for flash decoding on xpu#1118

Merged

IlyasMoutawwakil merged 11 commits intohuggingface:mainfrom kaixuanliu:flash_xpu

Feb 25, 2025

Commits on Jan 17, 2025

add support for flash decoding on xpu
kaixuanliu
committed

Commits on Jan 21, 2025

Merge branch 'main' into flash_xpu
kaixuanliu
authored
change block_size to 16 for xpu, as single_query_cached_kv_attention API does not support 64
kaixuanliu
committed

Commits on Jan 23, 2025

optimize the performance
kaixuanliu
committed

Commits on Feb 6, 2025

fix format issue
kaixuanliu
committed

Commits on Feb 17, 2025

refine code
kaixuanliu
committed

Commits on Feb 19, 2025

Commits on Feb 20, 2025

fix bug
kaixuanliu
committed
fix CI bug
kaixuanliu
committed