app.py

import gradio as gr
import spaces
from transformers import AutoModelForCausalLM, AutoTokenizer
from langchain.llms.base import LLM
from langchain.memory import ConversationBufferMemory
from langchain.chains import LLMChain, ConversationChain
from langchain.prompts import PromptTemplate, ChatPromptTemplate

@spaces.GPU
def initialize_model_and_tokenizer(model_name="KvrParaskevi/Llama-2-7b-Hotel-Booking-Model"):
    model = AutoModelForCausalLM.from_pretrained(model_name)
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    return model, tokenizer

model, tokenizer = initialize_model_and_tokenizer()

class CustomLLM(LLM):
    def _call(self, prompt, stop=None, run_manager=None) -> str:
        inputs = tokenizer(prompt, return_tensors="pt")
        result = model.generate(input_ids=inputs.input_ids, max_new_tokens=20)
        result = tokenizer.decode(result[0])
        return result

    @property
    def _llm_type(self) -> str:
        return "custom"

llm = CustomLLM()

template = """<<SYS>>
You are an AI having conversation with a human. Below is an instruction that describes a task. 
Write a response that appropriately completes the request.
Reply with the most helpful and logic answer. During the conversation you need to ask the user 
the following questions to complete the hotel booking task.
1) Where would you like to stay and when?
2) How many people are staying in the room?
3) Do you prefer any ammenities like breakfast included or gym?
4) What is your name, your email address and phone number? 
Make sure you receive a logical answer from the user from every question to complete the hotel 
booking process.
<</SYS>>
Previous conversation:
{history}
Human: {input}
AI:"""
prompt = PromptTemplate(template=template, input_variables=["history", "input"])

memory = ConversationBufferMemory(memory_key="history", llm = llm, prompt = prompt)
llm_chain = ConversationChain(prompt=prompt, llm=llm, memory = memory)

with gr.Blocks() as demo:
    chatbot = gr.Chatbot()
    msg = gr.Textbox()
    clear = gr.Button("Clear")
    #llm_chain, llm = init_chain(model, tokenizer)

    def user(user_message, history):
        return "", history + [[user_message, None]]
    @spaces.GPU
    def bot(history):
        print("Question: ", history[-1][0])
        bot_message = llm_chain.invoke(input=history[-1][0])
        print("Response: ", bot_message)
        history[-1][1] = ""
        history[-1][1] += bot_message
        return history

    msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(bot, chatbot, chatbot)
    clear.click(lambda: None, None, chatbot, queue=False)

demo.queue()
demo.launch()