Update app.py

app.py

@@ -19,7 +19,7 @@ def load_pipeline():
     pipe = pipeline("text-generation", 
                     model= model, 
                     tokenizer = tokenizer,
-                    max_new_tokens = 100,
+                    max_new_tokens = 50,
                     top_k = 30, 
                     early_stopping=True,
                     num_beams = 2,
@@ -37,8 +37,8 @@ def chat_interface(inputs):
         chat_history_tuples.append((message[0], message[1]))
     
     #result = llm_chain({"input": query, "history": chat_history_tuples})
-    result = llm_chain.predict(input = inputs)
-    return result
+    result = llm_chain.predict(input = inputs, max_length = 50)
+    return result["response"]
 
 llm = load_pipeline()
 chat_history = []
@@ -60,14 +60,14 @@ the following questions to complete the hotel booking task.
 Make sure you receive a logical answer from the user from every question to complete the hotel 
 booking process.
 <</SYS>>
-Previous conversation:
+Current conversation:
 {history}
 Human: {input}
 AI:"""
 prompt = PromptTemplate(template=template, input_variables=["history", "input"])
 
 memory = ConversationBufferMemory(memory_key="history", llm = llm)
-llm_chain = ConversationChain(llm=llm, memory = memory)
+llm_chain = ConversationChain(llm=llm, memory = memory, prompt= prompt)
 
 with gr.Blocks() as demo:
     #gr.Markdown("Hotel Booking Assistant Chat 🤗")