Datasets:

Modalities:
Text
Formats:
json
ArXiv:
Libraries:
Datasets
pandas
License:

"overall_image": null

#2
by lryyyy - opened

It is a super nice dataset, however in many datas the value of "overall_image" is empty. For example, in subset OBELICS/part00000/part00000.jsonl "id": 5755.
Can such issue be modified?

Multimodal Art Projection org

Thank you for pointing out the issue with the "overall_image" being empty. This indicates that there was an error in the compilation process when generating the overall image for that Markdown file, resulting in it being empty. I suggest either skipping that sample or attempting to compile it by yourself.
I plan to release a larger dataset in the near future to expand the existing data, but this issue may still persist. In discussions with the team members, we are considering publishing a data cleaning script or attempting to recompile the failed samples.
Thank you again for your feedback.

Thank you for pointing out the issue with the "overall_image" being empty. This indicates that there was an error in the compilation process when generating the overall image for that Markdown file, resulting in it being empty. I suggest either skipping that sample or attempting to compile it by yourself.
I plan to release a larger dataset in the near future to expand the existing data, but this issue may still persist. In discussions with the team members, we are considering publishing a data cleaning script or attempting to recompile the failed samples.
Thank you again for your feedback.

got it, thanks!

Sign up or log in to comment