use default cpu quantizer

manuelcandales · manuelcandales · commit 14ac9ffe7eec · 2024-12-05T14:27:12.000-05:00
diff --git a/examples/models/llama/source_transformation/quantize.py b/examples/models/llama/source_transformation/quantize.py
@@ -82,7 +82,6 @@ def quantize(  # noqa C901
 
         with torch.no_grad():
             model = UIntxWeightOnlyLinearQuantizer(
-                device="mps",
                 precision=torch.float32,
                 groupsize=group_size,
                 bitwidth=bitwidth
