[Refactor] Move fla to third party (#48500)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
@@ -10,12 +10,12 @@ import pytest
|
||||
import torch
|
||||
import torch.nn.functional as F
|
||||
|
||||
from vllm.model_executor.layers.fla.ops.kda import (
|
||||
from vllm.third_party.flash_linear_attention.ops.kda import (
|
||||
chunk_kda,
|
||||
chunk_kda_with_fused_gate,
|
||||
fused_kda_gate,
|
||||
)
|
||||
from vllm.model_executor.layers.fla.ops.l2norm import l2norm_fwd
|
||||
from vllm.third_party.flash_linear_attention.ops.l2norm import l2norm_fwd
|
||||
|
||||
DEVICE = "cuda"
|
||||
|
||||
|
||||
Reference in New Issue
Block a user