"olemeyer/docvqa-en-de-fr-es-it" is not available in huggingface anymore

[YanaSSS]

"olemeyer/docvqa-en-de-fr-es-it" dataset which is used in the notebook Creating_a_toy_DocVQA_dataset_for_Donut.ipynb is not available in huggingface anymore

[emigomez]

same issue, any update on this? there is another dataset available to use?

[marcomameli1992]

Can some someone explain how to create a personalized dataset for traini donut for DocVQA. I would like to test it on my own dataset and if it works I will release it as public available.

[NielsRogge]

Hi yes,

here's a guide: https://github.com/NielsRogge/Transformers-Tutorials/blob/master/Donut/DocVQA/Creating_a_toy_DocVQA_dataset_for_Donut.ipynb

[Ridwanlekan]

the reposity

"olemeyer/docvqa-en-de-fr-es-it" dataset which is used in the notebook Creating_a_toy_DocVQA_dataset_for_Donut.ipynb is not available in huggingface anymore

I have also checked. The repository doesnt exist. I also tried googling if the model exist, but couldnt find anything

[saeidMLK]

I have this issue too. Is there any dataset?

[Cherrysan]

Hi. Same issue here, but further into the tutorial you can see that Niels created a miniaturized version of the dataset as his toy set and pushed it to the hub. The path is "nielsr/docvqa_1200_examples". You can download this to check for format and then create your own.

[saeidMLK]

thank you for your response. yeah, I saw it. do you know how to create our own dataset from scratch? I want to create a dataset with my images. Could you tell me what the bounding_boxes column is?

[NielsRogge]

Creating a HF Dataset from scratch is explained here: https://huggingface.co/docs/datasets/image_dataset.