[CPU] layernorm & fused add-layernorm kernels (#14074)

This commit is contained in:
Zaili Wang
2025-12-12 08:58:23 +08:00
committed by GitHub
parent 9975acf50f
commit d6bd2d1126
5 changed files with 258 additions and 10 deletions

View File

@@ -361,7 +361,12 @@ class LayerNorm(CustomOp):
self,
x: torch.Tensor,
) -> torch.Tensor:
return self.forward_native(x)
if _is_cpu_amx_available:
return torch.ops.sgl_kernel.layernorm_cpu(
x, self.weight.data, self.variance_epsilon
)
else:
return self.forward_native(x)
class GemmaRMSNorm(CustomOp):