app.py

import os  
import huggingface_hub as hf_hub
import gradio as gr

client = hf_hub.InferenceClient(token = os.environ['HF_TOKEN'])
client.headers["x-use-cache"] = "0"

def image_interface(img):
    response = client.image_to_text(
        image = img, 
        model = 'Salesforce/blip-image-captioning-large'
    )

    return response
 
app = gr.Interface(
    fn = image_interface,
    inputs = gr.Image(type = 'filepath'),
    outputs = gr.Textbox(label = 'Image caption'),
    title = 'BLIP Image Captioning'
)


app.launch()