app.py

import torch
from transformers import IdeficsForVisionText2Text, AutoProcessor
from peft import PeftModel, PeftConfig
import gradio as gr

peft_model_id = "mrm8488/idefics-9b-ft-describe-diffusion-bf16-adapter"
device = "cuda" if torch.cuda.is_available() else "cpu"


config = PeftConfig.from_pretrained(peft_model_id)
model = IdeficsForVisionText2Text.from_pretrained(config.base_model_name_or_path, torch_dtype=torch.bfloat16)
model = PeftModel.from_pretrained(model, peft_model_id)
processor = AutoProcessor.from_pretrained(config.base_model_name_or_path)
model = model.to(device)
model.eval()

#Pre-determined best prompt for this fine-tune
prompt="Describe the following image:"

#Max generated tokens for your prompt
max_length=64

def predict(image):
    prompts = [[image, prompt]]
    inputs = processor(prompts[0], return_tensors="pt").to(device)
    generated_ids = model.generate(**inputs, max_length=max_length)
    generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
    generated_text = generated_text.replace(prompt,"")
    return generated_text

title = "Midjourney-like Image Captioning with IDEFICS"
description = "Gradio Demo for generating *Midjourney* like captions (describe functionality) with **IDEFICS**"

examples = [
    ["1_sTXgMwDUW0pk-1yK4iHYFw.png"],
    ["0_6as5rHi0sgG4W2Tq.png"],
    ["zoomout_2-1440x807.jpg"],
    ["inZdRVn7eafZNvaVre2iW1a538.webp"],
    ["cute-photos-of-cats-in-grass-1593184777.jpg"]
]
io = gr.Interface(fn=predict, 
                  inputs=[
                      gr.Image(label="Upload an image", type="pil"),
                  ],
                  outputs=[
                      gr.Textbox(label="IDEFICS Description")
                  ],
                  title=title, description=description, examples=examples,
                  allow_flagging=False, allow_screenshot=False)
io.launch(debug=True)