diff --git a/src/aiunn/upsampler/aiunn.py b/src/aiunn/upsampler/aiunn.py
index a5ba3b4..978a75d 100644
--- a/src/aiunn/upsampler/aiunn.py
+++ b/src/aiunn/upsampler/aiunn.py
@@ -17,24 +17,24 @@ class aiuNN(PreTrainedModel):
 
         # Enhanced approach
         scale_factor = self.config.upsample_scale
-        out_channels = self.base_model.config.num_channels * (scale_factor ** 2)
+        out_channels = self.aiia_model.config.num_channels * (scale_factor ** 2)
         self.pixel_shuffle_conv = nn.Conv2d(
-            in_channels=self.base_model.config.hidden_size,
+            in_channels=self.aiia_model.config.hidden_size,
             out_channels=out_channels,
-            kernel_size=self.base_model.config.kernel_size,
+            kernel_size=self.aiia_model.config.kernel_size,
             padding=1
         )
         self.pixel_shuffle = nn.PixelShuffle(scale_factor)
 
     def load_base_model(self, base_model: PreTrainedModel):
-        self.base_model = base_model
+        self.aiia_model = base_model
         
     def forward(self, x):
-        if self.base_model is None:
+        if self.aiia_model is None:
             raise ValueError("Base model is not loaded. Call 'load_base_model' before forwarding.")
             
         # Get base features - we need to extract the last hidden state if it's returned as part of a tuple/dict
-        base_output = self.base_model(x)
+        base_output = self.aiia_model(x)
         if isinstance(base_output, tuple):
             x = base_output[0]
         elif isinstance(base_output, dict):