apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-12-24 13:28:13 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

mv import (#5146)

Browse Source
This commit is contained in:
Ryan
2025-11-20 19:25:56 +08:00
committed by GitHub
parent c3994750b1
commit 0857099191
1 changed files with 1 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
fastdeploy/model_executor/layers/moe/fused_moe_wint2_backend.py
View File

@@ -19,6 +19,7 @@ from paddle import nn
import fastdeploy
from fastdeploy.distributed.communication import tensor_model_parallel_all_reduce
from fastdeploy.model_executor.ops.gpu import moe_expert_dispatch, moe_expert_reduce
from fastdeploy.utils import ceil_div
from ..quantization.quant_base import QuantMethodBase
@@ -266,7 +267,6 @@ class CutlassWint2FusedMoeMethod(Wint2MoeMethod):
Use Wint2 Triton Fusedmoe compute Fused MoE.
"""
gate_out = gate(x.cast("float32"))
from fastdeploy.model_executor.ops.gpu import moe_expert_dispatch
(
permute_input,
@@ -306,8 +306,6 @@ class CutlassWint2FusedMoeMethod(Wint2MoeMethod):
False,
)
from fastdeploy.model_executor.ops.gpu import moe_expert_reduce
fused_moe_out = moe_expert_reduce(
ffn_out,
topk_weights,