chatbot_bedrock.py

import os
#from langchain import PromptTemplate, HuggingFaceHub, LLMChain
from langchain.memory import ConversationBufferMemory
from langchain.chains import ConversationChain
import langchain.globals
from transformers import AutoModelForCausalLM, AutoTokenizer
import streamlit as st

@st.cache_resource 
def load_model():
    tokenizer = AutoTokenizer.from_pretrained("KvrParaskevi/Hotel-Assistant-Attempt4-Llama-2-7b")
    model = AutoModelForCausalLM.from_pretrained("KvrParaskevi/Hotel-Assistant-Attempt4-Llama-2-7b")
    return tokenizer,model


#Write function to connect to Bedrock
# def demo_chatbot():
#     # client = boto3.client('bedrock-runtime')
    
#     template = """Question: {question}

#     Answer: Let's think step by step."""
#     prompt = PromptTemplate(template=template, input_variables=["question"])
#     llm=HuggingFaceHub(repo_id="google/flan-t5-xl", model_kwargs={"temperature":1e-10})

#     question = "When was Google founded?"

#     print(llm_chain.run(question))
#     return demo_llm

#test out the code with the Predicgt method
    #return demo_llm.predict(input)
# = demo_chatbot('What is the temperature in Nuremberg today?')
#print(response)

def demo_miny_memory(model):
    # llm_data = get_Model(hugging_face_key)
    memory = ConversationBufferMemory(llm = model,max_token_limit = 512)
    return memory

def demo_chain(input_text, memory,model):
    # llm_data = get_Model(hugging_face_key)
    llm_conversation = ConversationChain(llm=model,memory=memory,verbose=langchain.globals.get_verbose())

    chat_reply = llm_conversation.predict(input=input_text)
    return chat_reply