import numpy as np


def l2_normalize(x):
    return x / np.linalg.norm(x, axis=-1, keepdims=True)


def main():
    embedding_dim = 1024
    haptic_embedding = l2_normalize(np.random.randn(1, embedding_dim).astype(np.float32))
    vision_embedding = l2_normalize(np.random.randn(1, embedding_dim).astype(np.float32))
    similarity = float((haptic_embedding * vision_embedding).sum())

    print("Cross-modal similarity:", similarity)


if __name__ == "__main__":
    main()