apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

weight only quant method support QKVGate_proj (#6612)

Browse Source
This commit is contained in:
chen
2026-03-03 11:19:32 +08:00
committed by GitHub
parent 3cc09418f1
commit 1cae7a0d53
1 changed files with 2 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/layers/quantization/weight_only.py
+2
View File
@@ -26,6 +26,7 @@ from fastdeploy import envs
from fastdeploy.model_executor.layers.linear import (
MergedColumnParallelLinear,
MergedReplicatedLinear,
QKVGateParallelLinear,
QKVParallelLinear,
)
from fastdeploy.model_executor.utils import (
@@ -252,6 +253,7 @@ class WeightOnlyLinearMethod(QuantMethodBase):
isinstance(layer, MergedColumnParallelLinear)
or isinstance(layer, QKVParallelLinear)
or isinstance(layer, MergedReplicatedLinear)
or isinstance(layer, QKVGateParallelLinear)
):
# Only MergedReplicatedLinear uses the default outdim.
tensor_output_dim = (self.model_format == "torch") ^ quant_attrs.get("output_dim", True)