Update app.py

app.py

@@ -8,8 +8,15 @@ model_path = hf_hub_download(
     repo_id="AstroMLab/AstroSage-8B-GGUF",
     filename="AstroSage-8B-Q8_0.gguf"
 )
-    
-    
+
+llm = Llama(
+    model_path=model_path,
+    n_ctx=2048,
+    chat_format="llama-3",
+    n_gpu_layers=-1,  # ensure all layers are on GPU
+    flash_attn=True,
+)
+
 # Placeholder responses for when context is empty
 GREETING_MESSAGES = [
     "Greetings! I am AstroSage, your guide to the cosmos. What would you like to explore today?",
@@ -26,15 +33,7 @@ def user(user_message, history):
 
 @spaces.GPU
 def bot(history):
-    """Initialize the bot and stream the response."""
-
-    llm = Llama(
-        model_path=model_path,
-        n_ctx=2048,
-        chat_format="llama-3",
-        n_gpu_layers=-1,  # ensure all layers are on GPU
-        flash_attn=True,
-    )
+    """Yield the chatbot response for streaming."""
     
     if not history:
         history = []
@@ -43,7 +42,7 @@ def bot(history):
     messages = [
         {
             "role": "system",
-            "content": "You are AstroSage, an intelligent AI assistant specializing in astronomy, astrophysics, and space science. You provide accurate, scientific information while making complex concepts accessible. You're enthusiastic about space exploration and maintain a sense of wonder about the cosmos."
+            "content": "You are AstroSage, an intelligent AI assistant specializing in astronomy, astrophysics, and cosmology. Provide accurate, scientific information while making complex concepts accessible. You're enthusiastic about space exploration and maintain a sense of wonder about the cosmos."
         }
     ]