I tried to make training in OCR-D using samples of german text from OCR-D repository.
I put only 20 tif/txt files in ground-truth folder and run make training with 1000 iterations.
I used one tif from ground-truth folder to test tesseract with trained data, and it didn't recognize any character.
What did I do wrong?