Ultra-compact vision-language model for document parsing
Try out Mistral's latest OCR with pdfs and images