OCR model for taking in images of documents as inputs and returning text as the output. It also supports non-latin characters.