Update app.py

app.py

@@ -49,7 +49,10 @@ embed_model = HuggingFaceInstructEmbeddings(
 model_name = "hkunlp/instructor-xl"
 model_kwargs = {'device': 'cpu'}
 encode_kwargs = {'normalize_embeddings': True}
-embed_model = InstructorEmbedding(
+
+from llama_index.core import Settings
+
+Settings.embed_model = InstructorEmbedding(
     model_name=model_name
 )
 
@@ -57,7 +60,7 @@ llm= LLM = G4FLLM(
         model=models.gpt_35_turbo_16k,
     )
 
-llm = LangChainLLM(llm=llm)
+Settings.llm = LangChainLLM(llm=llm)
 #embed_model=embed_model)
 
 service_context = ServiceContext.from_defaults(chunk_size=5512, llm=llm, embed_model=embed_model )