Update app.py

app.py

@@ -33,14 +33,19 @@ class ChatBot:
         # Generate a response using the model
         tokens = model.generate(
             inputs.to(model.device),
-            max_new_tokens=1024,
+            max_new_tokens=250,
             temperature=0.8,
-            do_sample=True
+            do_sample=False
         )
 
         # Decode and return the response
         response_text = tokenizer.decode(tokens[0], skip_special_tokens=False)
         return response_text
+        
+        # Free up memory
+        del tokens
+        torch.cuda.empty_cache()
+        return response_text
 
 bot = ChatBot()