[Refactor] Move fla to third party (#48500)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
@@ -286,7 +286,9 @@ class RMSNormGated(CustomOp):
|
||||
def forward_cuda(
|
||||
self, x: torch.Tensor, z: torch.Tensor | None = None
|
||||
) -> torch.Tensor:
|
||||
from vllm.model_executor.layers.fla.ops.layernorm_guard import rmsnorm_fn
|
||||
from vllm.third_party.flash_linear_attention.ops.layernorm_guard import (
|
||||
rmsnorm_fn,
|
||||
)
|
||||
|
||||
return rmsnorm_fn(
|
||||
x,
|
||||
|
||||
Reference in New Issue
Block a user