From 6dd5e52823cc0ca8ddc9c4377d29ead37cc09a95 Mon Sep 17 00:00:00 2001
From: Kyle Sayers <kylesayrs@gmail.com>
Date: Mon, 3 Feb 2025 16:29:56 -0500
Subject: [PATCH] Squelch MLA warning for Compressed-Tensors Models (#12704)

Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>
---
 vllm/config.py | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/vllm/config.py b/vllm/config.py
index d70a637956edf..2f4a7ad769d98 100644
--- a/vllm/config.py
+++ b/vllm/config.py
@@ -986,6 +986,9 @@ class ModelConfig:
 
     @property
     def use_mla(self) -> bool:
+        if not self.is_deepseek_mla or envs.VLLM_MLA_DISABLE:
+            return False
+
         if self.quantization is not None and self.quantization not in [\
             "fp8", "compressed-tensors"]:
             logger.warning(
@@ -1012,8 +1015,7 @@ class ModelConfig:
                         quant_config)
                     return False
 
-        use_mla = (self.is_deepseek_mla and not envs.VLLM_MLA_DISABLE)
-        return use_mla
+        return True
 
     @property
     def supported_runner_types(self) -> Set[RunnerType]: