correct defect

app.py

@@ -65,7 +65,13 @@ import langchain
 import asyncio
 from playwright.async_api import async_playwright
 
-os.environ['MPLCONFIGDIR'] = os.path.expanduser('/home/user/matplotlib_cache')
+
+from langchain.embeddings.huggingface import HuggingFaceEmbeddings
+
+
+
+
+# os.environ['MPLCONFIGDIR'] = os.path.expanduser('/home/user/matplotlib_cache')
 
 async def start_playwright(question: str):
     pw = await async_playwright().start()
@@ -489,6 +495,12 @@ chat = AzureChatOpenAI(
     deployment_name=os.environ["deployment_name"],
     temperature=0,
 )
+
+
+
+
+
+
 llm = chat
 
 llm_math = LLMMathChain.from_llm(llm)
@@ -507,7 +519,22 @@ tools2 = [DB_Search2(), duckduckgo_tool2, wikipedia_tool2, python_tool2, math_to
 
 # tools = load_tools(["Vector Database Search","Wikipedia Search","Python REPL","llm-math"], llm=llm)
 
-embeddings = OpenAIEmbeddings(deployment="model_embedding", chunk_size=15)
+# Openai embedding
+embeddings_openai = OpenAIEmbeddings(deployment="model_embedding", chunk_size=15)
+
+# huggingface embedding model
+embed_model_id = 'sentence-transformers/all-MiniLM-L6-v2'
+
+# device = f'cuda:{cuda.current_device()}' if cuda.is_available() else 'cpu'
+device = 'cpu'
+embed_model = HuggingFaceEmbeddings(
+    model_name=embed_model_id,
+    model_kwargs={'device': device},
+    encode_kwargs={'device': device, 'batch_size': 32}
+)
+
+embeddings = embed_model
+
 
 # embeddings = OpenAIEmbeddings(deployment="model_embedding_2", chunk_size=15)
 
@@ -1119,7 +1146,7 @@ with gr.Blocks() as demo:
 
     with gr.Column() as main2:
         title = gr.Markdown("""# <center> STLA BABY - YOUR FRIENDLY GUIDE
-                            <center> v0.5: Powered by MECH Core Team"""),
+                            <center> v0.6: Powered by MECH Core Team - GPT4 REMOTE MODE"""),
         chatbot = gr.Chatbot()
         with gr.Row():
             inputtext = gr.Textbox(