apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 08:37:06 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[fix] w4a8 model loading and hadamard config (#3013)

Browse Source
This commit is contained in:
Yuan Xiaolan
2025-07-28 18:17:59 +08:00
committed by GitHub
parent c8bf8b3913
commit b1d787a272
4 changed files with 32 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
fastdeploy/model_executor/layers/moe/moe.py
View File

@@ -19,6 +19,9 @@ from paddle import nn
from paddleformers.utils.log import logger
from fastdeploy import envs
from fastdeploy.model_executor.layers.moe.fused_moe_cutlass_backend import (
CutlassW4A8MoEMethod,
)
from fastdeploy.model_executor.layers.utils import get_tensor
from fastdeploy.worker.experts_manager import RedundantExpertManger
@@ -385,7 +388,10 @@ class FusedMoE(nn.Layer):
self.gate_weight.set_value(gate_weight_tensor.astype("float32"))
if self.fd_config.model_config.is_quantized:
self.quant_method.process_prequanted_weights(self, state_dict)
if isinstance(self.quant_method, CutlassW4A8MoEMethod):
self.quant_method.create_weights(self, state_dict)
else:
self.quant_method.process_prequanted_weights(self, state_dict)
else:
self.quant_method.create_weights(self, state_dict)