ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-03 05:01:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

Reduce redundant space in bias tensor (#2024)

Browse Source 
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>
This commit is contained in:
rocking
2025-03-28 21:58:06 +08:00
committed by GitHub
parent a82f338fb9
commit 8a20b62e91
2 changed files with 7 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
include/ck_tile/ops/fmha/kernel/fmha_fwd_kernel.hpp
View File

@@ -983,7 +983,7 @@ struct FmhaFwdKernel
}
if constexpr(BiasEnum == BlockAttentionBiasEnum::ELEMENTWISE_BIAS)
{
batch_offset_bias = query_start * kargs.stride_bias + key_start;
batch_offset_bias = query_start * kargs.stride_bias;
}
if constexpr(kStoreLSE)
{