Document processing for PDFs and scanned documents
ljvmiranda921 opened this issue · comments
Lj Miranda commented
https://github.com/mindee/doctr
- Maybe Prodigy + PDF recipe?
- Train a model straight from Prodigy?
Lj Miranda commented
This evolved into this issue (as handled by this PR: #304). I'm not using doctr anymore, but Hugging Face's LayoutLMv3 model.