Update app.py

app.py

@@ -1,31 +1,23 @@
 import streamlit as st
-from langchain.llms import HuggingFaceHub
+from transformers import pipeline
 
-# Function to return the response
-def generate_answer(query):
-    llm = HuggingFaceHub(
-        repo_id="huggingfaceh4/zephyr-7b-alpha",
-        model_kwargs={"temperature": 0.5, "max_length": 64, "max_new_tokens": 512}
-    )
-    prompt = f"""
-        You are helpful doctor assistant that gives patients advices. Please answer the patients in a kind and helpful way.
-        </s>
-        {query}</s>
-    """
-    result = llm.predict(prompt)
-    return result
+# Load the conversational pipeline
+conversational_pipeline = pipeline("conversational")
 
-# App UI starts here
-st.set_page_config(page_title="LangChain Demo", page_icon=":robot:")
-st.header("LangChain Demo")
+# Streamlit app header
+st.set_page_config(page_title="Conversational Model Demo", page_icon="🤖")
+st.header("Conversational Model Demo")
 
-# Gets User Input
-user_input = st.text_input("You: ", key="input")
+# Input for user message
+user_message = st.text_input("You:", "")
 
-submit = st.button("Generate")
+if st.button("Send"):
+    # Format the conversation for the conversational pipeline
+    conversation_history = [{"role": "system", "content": "You are an AI assistant."},
+                           {"role": "user", "content": user_message}]
 
-# If the button clicked
-if submit:
-    st.subheader("Answer:")
-    response = generate_answer(user_input)
-    st.write(response)
+    # Get the model's response
+    model_response = conversational_pipeline(conversation_history)[0]['generated_responses'][0]
+
+    # Display the model's response
+    st.text_area("Model:", model_response, height=100)