app.py

import streamlit as st
from transformers import AutoModel, AutoTokenizer
import torch

model_name = "your-model-name-here"
model = AutoModel.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

st.title("Multimodal Model Demo")

text_input = st.text_input("Enter text prompt (optional)")
image_input = st.file_uploader("Upload image (optional)", type=["jpg", "jpeg", "png"])

if text_input or image_input:
    inputs = tokenizer(text_input, image_input, return_tensors="pt", padding=True)
    outputs = model(**inputs)
    output_text = tokenizer.decode(outputs.logits[0], skip_special_tokens=True)

    st.write("Output:")
    st.write(output_text)