app.py

from transformers import (
    Blip2VisionConfig, 
    Blip2QFormerConfig, 
    OPTConfig, 
    Blip2Config, 
    Blip2ForConditionalGeneration, 
    Blip2VisionModel, 
    Blip2Processor,
    AutoProcessor
)
from PIL import Image
import requests
import torch
import gradio as gr

config = Blip2Config()
model = Blip2ForConditionalGeneration(config)
config = model.config

vis_config = Blip2VisionConfig()
model = Blip2VisionModel(vis_config)
config_2 = model.config

processor = AutoProcessor.from_pretrained('Salesforce/blip-image-captioning-large')
model = Blip2ForConditionalGeneration.from_pretrained('Salesforce/blip-image-captioning-large')

def captioning(image):
    inputs = processor(images=image, return_tensors='pt')
    generated_ids = model.generate(**inputs)
    generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
    return image, generated_text

demo = gr.Interface(
    captioning,
    inputs=gr.Image(type="pil"),
    outputs = ['image', 'text']
)

demo.launch()