Update app.py

app.py

@@ -38,7 +38,7 @@ def messages_to_prompt(messages):
 
 def load_RAG_pipeline():
     # LLM
-    quantization_config = {}  # dirty fix for CPU/GPU support
+    quantization_config = None # dirty fix for CPU/GPU support
     if torch.cuda.is_available():
         from transformers import BitsAndBytesConfig