add cache dir

src/app.py

@@ -21,6 +21,13 @@ def load_model(from_pretrained, base_model, cache_dir, device):
     Load and cache the model to avoid reloading on every user interaction.
     This function runs only once when the app starts or when parameters change.
     """
+    is_hf_space = os.environ.get('SPACE_ID') is not None
+    if is_hf_space:
+        cache_dir = None
+        device = 'cpu'
+        print("Using **CPU** now!")
+    else:
+        cache_dir = cache_dir
     with st.spinner("🔄 Loading model... This may take a moment on first launch."):
         model = ComputeScore.from_pretrained(
             from_pretrained,