fix: update ONNX model cache handling in inference functions

app.py

@@ -93,7 +93,7 @@ _onnx_model_cache = {}
 
 def infer_onnx_model(hf_model_id, preprocessed_image_np, model_config: dict):
     try:
-        ort_session, _, _ = get_onnx_model_from_cache(hf_model_id)
+        ort_session, _, _ = get_onnx_model_from_cache(hf_model_id, _onnx_model_cache, load_onnx_model_and_preprocessor)
         
         # Debug: Print expected input shape from ONNX model
         for input_meta in ort_session.get_inputs():
@@ -134,7 +134,9 @@ class ONNXModelWrapper:
 
     def load(self):
         if self._session is None:
-            self._session, self._preprocessor_config, self._model_config = get_onnx_model_from_cache(self.hf_model_id)
+            self._session, self._preprocessor_config, self._model_config = get_onnx_model_from_cache(
+                self.hf_model_id, _onnx_model_cache, load_onnx_model_and_preprocessor
+            )
             logger.info(f"ONNX model {self.hf_model_id} loaded into wrapper.")
 
     def __call__(self, image_np):