Change warning logs to debug for unimplemented MXFP4 Linear/Attention (#29441)

Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>

Author: mgoin

vllm/model_executor/layers/quantization/mxfp4.py

@@ -196,9 +196,10 @@ def get_quant_method(
             # TODO: Add support for MXFP4 Linear Method.
             # MXFP4 LinearMethod is available in AMD-Quark, refer to that implementation
             # if you are interested in enabling MXFP4 here.
-            logger.warning_once(
+            logger.debug_once(
                 "MXFP4 linear layer is not implemented - falling back to "
-                "UnquantizedLinearMethod."
+                "UnquantizedLinearMethod.",
+                scope="local",
             )
             return UnquantizedLinearMethod()
         elif isinstance(layer, FusedMoE):
@@ -208,9 +209,10 @@ def get_quant_method(
                 return Mxfp4MoEMethod(layer.moe_config)
         elif isinstance(layer, Attention):
             # TODO: Add support for MXFP4 Attention.
-            logger.warning_once(
+            logger.debug_once(
                 "MXFP4 attention layer is not implemented. "
-                "Skipping quantization for this layer."
+                "Skipping quantization for this layer.",
+                scope="local",
             )
         return None