Revert "Switch to new ao quant api for 8da4w (#8501)"

This reverts commit f3fc096.

Author: swolchok

examples/models/llama/source_transformation/quantize.py

@@ -119,10 +119,11 @@ def quantize(  # noqa C901
         # Check for required args
         if group_size is None:
             raise Exception("For 8da4w quantization, group size must be specified.")
+        from torchao.quantization.quant_api import Int8DynActInt4WeightQuantizer
 
-        from torchao.quantization import int8_dynamic_activation_int4_weight, quantize_
-
-        quantize_(model, int8_dynamic_activation_int4_weight(group_size=group_size))
+        model = Int8DynActInt4WeightQuantizer(
+            precision=torch_dtype, groupsize=group_size
+        ).quantize(model)
 
         if verbose:
             print("quantized model:", model)
@@ -662,7 +663,7 @@ def convert_for_runtime(self) -> nn.Module:
     def quantized_model(self) -> nn.Module:
         model_updated_state_dict = self.create_quantized_state_dict(self.packed)
         self.convert_for_runtime()
-        self.mod.load_state_dict(model_updated_state_dict, assign=True)
+        self.mod.load_state_dict(model_updated_state_dict)
         return self.mod