Update app.py

app.py

@@ -20,12 +20,17 @@ if "scrape_results" not in st.session_state:
 # Load Zephyr model
 @st.cache_resource
 def load_model():
-    return pipeline(
-        "text-generation",
-        model="HuggingFaceH4/zephyr-7b-beta",
-        torch_dtype=torch.float16,
-        device_map="auto",
-    )
+    try:
+        return pipeline(
+            "text-generation",
+            model="HuggingFaceH4/zephyr-7b-beta",
+            torch_dtype=torch.float16,
+            device_map="auto",
+            model_kwargs={"load_in_8bit": True}  # Use 8-bit quantization to reduce memory usage
+        )
+    except Exception as e:
+        st.error(f"Error loading model: {str(e)}")
+        return None
 
 # Initialize the model
 model = load_model()