identify whether pdf has hand written text or check boxes, its flattened pdf, so no form elements, without LLM, is it possible to do? Main use case is I need to identify first and if it has hand written text or check boxes, then only I want to pass to multimodal LLM in order to save costs
No need for llm
You can implement an algorithm yourself to do it yourself
how do i do it then 🤔
You can use densenet neural architecture though
Обсуждают сегодня