corrupted text

[xudesheng]

Use this PDF as an example: https://arxiv.org/pdf/2407.01481

"efficient" and "difficult" are both corrupted in the same pattern: "ffi" was recognized as HEX "EF 81 8E."

I don't see the same issue if I use the pdfium library to grab the text. "ffi" is just one example.

speed-up and ecient use of resources is essential

One of the more dicult aspects of High Performance Computing