app.py

import gradio as gr
from transformers import AutoModel, AutoProcessor
from PIL import Image
import torch

model_name_or_path = "lyttt/VLV_captioner"
model = AutoModel.from_pretrained(model_name_or_path, revision="master", trust_remote_code=True,low_cpu_mem_usage=False)

def greet(image):
    image = image.convert("RGB") 
    with torch.no_grad():
        outputs = model([image]).generated_text[0]
    return output

demo = gr.Interface(fn=greet, inputs="image", outputs="text")
demo.launch()