app.py

import gradio as gr
import numpy as np
from PIL import Image
from transformers import AutoProcessor, AutoModelForPreTraining
import os

# PaliGemma settings
access_token = os.getenv('HF_TOKEN')
processor = AutoProcessor.from_pretrained("google/paligemma-3b-mix-224", token=access_token)
model = AutoModelForPreTraining.from_pretrained("google/paligemma-3b-mix-224", token=access_token)

def response_request(image,prompt):
    inputs = processor(prompt, image, return_tensors="pt")
    output = model.generate(**inputs, max_new_tokens=100, do_sample=False)
    response = processor.decode(output[0], skip_special_tokens=True)[len(prompt):]
    return response

# Interface
iface = gr.Interface(
    fn=response_request, 
    inputs=[
        gr.Image(type="pil", label="Upload Image"),
        gr.Textbox(label="Prompt")
    ],
    outputs=[
        gr.Textbox(label="Response")
    ],
    title="PaliGemma (google/paligemma-3b-pt-896)"
)

iface.launch()