Update app.py

app.py

@@ -5,6 +5,7 @@ from langchain.llms import HuggingFacePipeline
 from langchain import PromptTemplate
 from langchain.embeddings import HuggingFaceEmbeddings
 from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
 from langchain.vectorstores import Chroma
 from langchain.document_loaders import (
     CSVLoader,
@@ -109,7 +110,8 @@ def load_model():
     #     # temperature=temperature,  # type: ignore
     # )
     llm = CTransformers(
-        model="TheBloke/Llama-2-7B-Chat-GGML"
+        model="TheBloke/Llama-2-7B-Chat-GGML",
+        callbacks=[StreamingStdOutCallbackHandler()]
         # model_type=model_type,
         # max_new_tokens=max_new_tokens,  # type: ignore
         # temperature=temperature,  # type: ignore