diff --git a/server/text_generation_server/layers/awq/quantize/ipex.py b/server/text_generation_server/layers/awq/quantize/ipex.py
index 84cd7a21..842e9623 100644
--- a/server/text_generation_server/layers/awq/quantize/ipex.py
+++ b/server/text_generation_server/layers/awq/quantize/ipex.py
@@ -44,5 +44,4 @@ class WQLinear(nn.Module):
     def forward(self, x):
         out_shape = x.shape[:-1] + (self.out_features,)
         out = self.woq_linear(x.reshape(-1, x.shape[-1]))
-        out = out + self.bias if self.bias is not None else out
         return out.reshape(out_shape)
diff --git a/server/text_generation_server/layers/gptq/ipex.py b/server/text_generation_server/layers/gptq/ipex.py
index ab9c9e24..48584e90 100644
--- a/server/text_generation_server/layers/gptq/ipex.py
+++ b/server/text_generation_server/layers/gptq/ipex.py
@@ -122,5 +122,4 @@ class QuantLinear(nn.Module):
     def forward(self, x):
         out_shape = x.shape[:-1] + (self.outfeatures,)
         out = self.woq_linear(x.reshape(-1, x.shape[-1]))
-        out = out + self.bias if self.bias is not None else out
         return out.reshape(out_shape)