videoinput/hf.py

import requests
import torch
from PIL import Image
from io import BytesIO

from transformers import AutoProcessor, AutoModelForVision2Seq
from transformers.image_utils import load_image

DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")

image1 = load_image(
    "https://cdn.pixabay.com/photo/2015/04/23/22/00/tree-736885_960_720.jpg"
)

processor = AutoProcessor.from_pretrained("HuggingFaceM4/idefics2-8b")
model = AutoModelForVision2Seq.from_pretrained("HuggingFaceM4/idefics2-8b").to(DEVICE)