[BugFix][TritonMLA] Process weights after model loading for GGUF (vllm-project#14555)

Signed-off-by: TianyuanWu <Tianyuan.Wu@amd.com>
Signed-off-by: Mu Huai <tianbowen.tbw@antgroup.com>

Author: tywuAMD

vllm/model_executor/model_loader/loader.py

@@ -1330,11 +1330,14 @@ def load_model(self, vllm_config: VllmConfig) -> nn.Module:
                 local_model_path, gguf_weights_map):
             model_config.hf_config.update({"tie_word_embeddings": True})
 
+        target_device = torch.device(device_config.device)
         with set_default_torch_dtype(model_config.dtype):
-            with torch.device(device_config.device):
+            with target_device:
                 model = _initialize_model(vllm_config=vllm_config)
             model.load_weights(
                 self._get_weights_iterator(local_model_path, gguf_weights_map))
+
+            _process_weights_after_loading(model, model_config, target_device)
         return model