[MoE] Move FlashInfer CuteDSL experts into fused_moe/experts/ (#37759)
Signed-off-by: Robert Shaw <robertgshaw2@gmail.com>
This commit is contained in:
@@ -17,7 +17,7 @@ from flashinfer import fp4_quantize
|
|||||||
from torch.nn import functional as F
|
from torch.nn import functional as F
|
||||||
|
|
||||||
from vllm.model_executor.layers.activation import SiluAndMul
|
from vllm.model_executor.layers.activation import SiluAndMul
|
||||||
from vllm.model_executor.layers.fused_moe.flashinfer_cutedsl_moe import (
|
from vllm.model_executor.layers.fused_moe.experts.flashinfer_cutedsl_moe import (
|
||||||
flashinfer_cutedsl_moe_masked,
|
flashinfer_cutedsl_moe_masked,
|
||||||
)
|
)
|
||||||
from vllm.utils.flashinfer import (
|
from vllm.utils.flashinfer import (
|
||||||
|
|||||||
@@ -86,7 +86,7 @@ def backend_to_kernel_cls(
|
|||||||
return [FlashInferExperts]
|
return [FlashInferExperts]
|
||||||
|
|
||||||
elif backend == NvFp4MoeBackend.FLASHINFER_CUTEDSL:
|
elif backend == NvFp4MoeBackend.FLASHINFER_CUTEDSL:
|
||||||
from vllm.model_executor.layers.fused_moe.flashinfer_cutedsl_moe import (
|
from vllm.model_executor.layers.fused_moe.experts.flashinfer_cutedsl_moe import ( # noqa: E501
|
||||||
FlashInferCuteDSLExperts,
|
FlashInferCuteDSLExperts,
|
||||||
)
|
)
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user