sitloboi2012/CMDS_Multimodal_Document
Dataset Card for Cyrillic Multimodel Document (CMDS) This is the dataset consists of 3789 pairs of images and text across 31 categories downloaded from the Bulgarian ministry of finance Dataset Summary This dataset card aims to be a base template for new datasets. It has been generated using this raw template. Supported Tasks and Leaderboards Uses this dataset for downstream task like Document Classification, Image Classification or Text… See the full description on the dataset page: https://huggingface.co/datasets/sitloboi2012/CMDS_Multimodal_Document.
Dataset Card for Cyrillic Multimodel Document (CMDS)
This is the dataset consists of 3789 pairs of images and text across 31 categories downloaded from the Bulgarian ministry of finance
Dataset Summary
This dataset card aims to be a base template for new datasets. It has been generated using this raw template.
Supported Tasks and Leaderboards
Uses this dataset for downstream task like Document Classification, Image Classification or Text Classification (Sequences Classification). Suitable for multimodal Model like LayoutLm Family, Donut, etc.
Languages
Bulgarian
Data Fields
- _text_ (bytes): the text appear in the document
- _filename_ (str): the name of the file
- _image_ (PIL.Image): the image of the document
- _label_ (str): the label of the document. There are 31 differences labels