Implement current SOTA model for image text matching task - View it on GitHub