A Specialized Vision-Language Model (VLM) designed for financial document analysis and question answering