8 ms·LLMs don't offer much value for our use case, almost all values are just numbersby is_true 2y agoLLMs don't offer much value for our use case, almost all values are just numbersinfecto 2y agoThen you should be using something like Textract or other tooling in that space. Multimodal LLMs are no replacement.is_true 2y agoWe use opencv + tesseract and easyocr