app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer
from inference import get_bot_response
from rag import get_context
from config import config
from huggingface_hub import InferenceClient


model_name = "mistralai/Mistral-7B-Instruct-v0.2"


client = InferenceClient(model_name)
print("tokenizer start loading")
tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
print("tokenizer loaded")
print("model start loading")
model = AutoModelForCausalLM.from_pretrained(model_name,
                                             device_map="auto",
                                             trust_remote_code=False,
                                             revision="main")
print("model loaded")

# model = AutoModelForCausalLM.from_pretrained(config["model_checkpoint"],
#                                                 device_map="auto",
#                                                 trust_remote_code=False,
#                                                 revision="main")


def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages = [{"role": "system", "content": system_message}]

    request = message
    context = get_context(request, config["top_k"])
    response = get_bot_response(request, context, model, tokenizer)

    return response

demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
)


if __name__ == "__main__":
    demo.launch()