app.py

import gradio as gr
from huggingface_hub import InferenceClient, login
import random
from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint, HuggingFacePipeline
from langchain.schema import AIMessage, HumanMessage
import os

login(token=os.environ["HUGGINGFACEHUB_API_TOKEN"])

llm = HuggingFaceEndpoint(
    repo_id="HuggingFaceH4/zephyr-7b-beta",
    task="text-generation",
    max_new_tokens=512,
    do_sample=False,
    repetition_penalty=1.03,
)

model = ChatHuggingFace(llm=llm)

def predict(message, history):
    history_langchain_format = []
    for msg in history:
        if msg['role'] == "user":
            history_langchain_format.append(HumanMessage(content=msg['content']))
        elif msg['role'] == "assistant":
            history_langchain_format.append(AIMessage(content=msg['content']))
    history_langchain_format.append(HumanMessage(content=message))
    gpt_response = model.invoke(history_langchain_format)
    return gpt_response.content

demo = gr.ChatInterface(
    predict,
    type="messages"
)

demo.launch()