Question: Training seems to work fine, but using traineddata file produces garbage

[lzhaxi]

Archive.zip

Uploaded my training log and .traineddata files along with a sample image. Log seems to indicate that the model is correctly getting the text, but if I try running it myself on the exact same image that is being used for evaluation, I get complete garbage output. I'm guessing that the .traineddata file is not the issue, but somehow tesseract is not reading it correctly, but not sure how to fix. Would appreciate any help

[lzhaxi]

Bump: attached my ground-truth files if anyone would like to try to reproduce this, maybe somehow the training has an issue scode-ground-truth.zip

[zdenop]

Did run training from scratch or based on already existing model? Please provide all steps with logs how did you run training.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.