# app.py
import gradio as gr
from transformers import AutoProcessor, AutoModelForVision2Seq
from PIL import Image
import torch

MODEL_ID = "Qwen/Qwen2-VL-7B-Instruct"
device = "cuda" if torch.cuda.is_available() else "cpu"

print(f"Loading model: {MODEL_ID}")
processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForVision2Seq.from_pretrained(
    MODEL_ID,
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
    device_map="auto"
)

def describe_ingredients(image):
    """Ask the model to identify ingredients in an image."""
    prompt = "Describe the food in this image and list its likely ingredients."
    inputs = processor(images=image, text=prompt, return_tensors="pt").to(device)

    output = model.generate(**inputs, max_new_tokens=128)
    result = processor.batch_decode(output, skip_special_tokens=True)[0]
    return result.strip()

demo = gr.Interface(
    fn=describe_ingredients,
    inputs=gr.Image(type="pil"),
    outputs="text",
    title="🍱 Food Ingredient Analyzer (Qwen2-VL)",
    description="Upload a food image — the model will describe it and list possible ingredients."
)

if __name__ == "__main__":
    demo.launch()