[Feature][Quantization] auto_round support for mixed bits quantization (#23812)

Signed-off-by: n1ck-guo <heng.guo@intel.com>
Signed-off-by: Heng Guo <heng.guo@intel.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

vllm/model_executor/layers/quantization/auto_round.py

@@ -436,6 +436,12 @@ class AutoRoundConfig(QuantizationConfig):
             return None
 
     def get_quant_method(self, layer: torch.nn.Module, prefix: str):
+        if prefix and self.extra_config:
+            for layer_name in self.extra_config:
+                if (
+                    layer_name == prefix or layer_name == f"model.{prefix}"
+                ) and self.extra_config[layer_name].get("bits", 16) >= 16:
+                    return UnquantizedLinearMethod()
         if (
             current_platform.is_cpu()
             or current_platform.is_xpu()