from PIL import Image
import requests
from transformers import Blip2Processor, Blip2Model, Blip2ForConditionalGeneration
import torch

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

processor = Blip2Processor.from_pretrained("Salesforce/blip2-flan-t5-xxl", load_in_8bit=True, device_map="auto")

model = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-flan-t5-xxl", load_in_8bit=True, device_map="auto")

url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)

inputs = processor(image,  return_tensors="pt").to(device, torch.float16)

out = model.generate(**inputs, max_length=64, min_length=20)
print(i,': ',processor.decode(out[0], skip_special_tokens=True))