app.py

import streamlit as st
from PIL import Image

from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
model = AutoModelForImageTextToText.from_pretrained("Salesforce/blip-image-captioning-large")

enable = st.checkbox("Enable camera")
picture = st.camera_input("Take a picture", disabled=not enable)

if picture:
	raw_image = Image.open(picture)

	# conditional image captioning
	text = "A view of a person in"
	inputs = processor(raw_image, text, return_tensors="pt")

	out = model.generate(**inputs)
	st.write(processor.decode(out[0], skip_special_tokens=True))