### **📄 inference.py (optional CLI tool)**

# inference.py
import argparse
import torch
from model import BiomedClipClassifier, predict_from_paths

def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("--weights", type=str, default=".")
    parser.add_argument("--mri", type=str, required=True, help="Path to NIfTI MRI file")
    parser.add_argument("--text", type=str, required=True, help="Clinical text")
    args = parser.parse_args()

    device = "cuda" if torch.cuda.is_available() else "cpu"
    model = BiomedClipClassifier.from_pretrained(args.weights, device=device)

    pred, probs = predict_from_paths(model, args.mri, args.text, device=device)
    print("Prediction:", pred)
    print("Probabilities [CN, MCI, Dementia]:", [round(p, 4) for p in probs])

if __name__ == "__main__":
    main()