Applied quantization for linear with bias=True in pre_quantization

Differential Revision: D71573144

Pull Request resolved: #9472

Author: YIWENX14

examples/models/llama/source_transformation/pre_quantization.py

@@ -44,7 +44,7 @@ def replacement_fn(child: torch.nn.Module) -> torch.nn.Module:
             # pyre-fixme[6]: For 2nd argument expected `int` but got `Union[Module,
             #  Tensor]`.
             child.out_features,
-            bias=False,
+            bias=child.bias is not None,
             device=child.weight.device,
             groupsize=group_size,
             precision=precision,