first version of pdftoimage and ocr

binary mode consumes a lot of RAM, added non binary mode

Merge branch 'main' of https://huggingface.co/datasets/jordyvl/DUDE_loader into main

Dataloader supporting trainval - remaining memory issue

Update data/DUDE_dataset-sample_gt.json

Upload DUDE_train-val-test_binaries.tar.gz

Releasing the binaries to the public - GT to follow

Upload 2 files

close to done for bboxes

Merge branch 'main' of https://huggingface.co/datasets/jordyvl/DUDE_loader into main

updated loader for testing different generate examples approach

Rename dude_loader.py to DUDE_loader.py

renamed and quention's iteration suggestion

renamed and quention's iteration suggestion

start of repo

initial commit