from transformers import  BlipForConditionalGeneration
from transformers import AutoProcessor
from PIL import Image
import requests
import gradio as gr

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


processor = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")


image = processor(image, return_tensors="pt")
generated_ids = model.generate(**image)
generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
print(generated_text)



def launch(input):
    url = input
    image = Image.open(requests.get(url, stream=True).raw).convert("RGB")
    processor = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
    model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
    image = processor(image, return_tensors="pt")
    generated_ids = model.generate(**image)
    generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
    return generated_text
  
  
iface = gr.Interface(fn=launch, inputs="text", outputs="text")
  
iface.launch()