diff options
| author | Gunes Bayir <gunes.bayir@arm.com> | 2024-04-09 23:13:04 +0100 |
|---|---|---|
| committer | Gunes Bayir <gunes.bayir@arm.com> | 2024-04-11 12:58:45 +0000 |
| commit | cfca87b91def4f455630f2094447dc0500b6256c (patch) | |
| tree | 9985ca8ad1910d48a84aa9781fe3156e614ff5f4 /src/cpu/kernels/softmax/list.h | |
| parent | f1f1f87132690a8061801ef1a4638d637c780df7 (diff) | |
| download | ComputeLibrary-cfca87b91def4f455630f2094447dc0500b6256c.tar.gz | |
Add SME2 implementation of softmax for FP16
In addition to the softmax kernel, this patch fixes minor issues in the fp32 implementation.
Resolves: COMPMID-6920
Change-Id: Ibbd9f0af5f2a93fba0e92d72ba437279c34149d3
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11402
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/cpu/kernels/softmax/list.h')
| -rw-r--r-- | src/cpu/kernels/softmax/list.h | 3 |
1 files changed, 3 insertions, 0 deletions
diff --git a/src/cpu/kernels/softmax/list.h b/src/cpu/kernels/softmax/list.h index 16fbd31a19..1bb8ed50f0 100644 --- a/src/cpu/kernels/softmax/list.h +++ b/src/cpu/kernels/softmax/list.h @@ -42,6 +42,9 @@ DECLARE_SOFTMAX_KERNEL(neon_qasymm8_signed_softmax); void sme2_fp32_softmax( const ITensor *in, void *const tmp, ITensor *out, const float beta, int axis, const Window &window); +void sme2_fp16_softmax( + const ITensor *in, void *const tmp, ITensor *out, const float beta, int axis, const Window &window); + #endif // ARM_COMPUTE_ENABLE_SME2 #undef DECLARE_SOFTMAX_KERNEL |