[HELP] Need training dataset and configuration files

[quoclam-doit]

Hi @arvindrajan92,

Thanks for sharing this amazing project! I'm currently trying to reproduce your work but I'm missing the training dataset and training scripts.

Could you please help me with:

The dataset used for training (train.uttlist, validation.uttlist, test.uttlist)
Training configuration/scripts
Any preprocessing steps needed
I understand if you can't share everything, but any guidance would be greatly appreciated! 🙏

If you prefer to discuss this privately or share sensitive resources, feel free to reach out via email: lyquoclam123@gmail.com

Thanks in advance!

[github-actions]

Hi @{{ sender.login }}! 👋

Welcome to DTrOCR project! Thank you for your first issue. We're excited to have you here and will respond at the soonest. Happy coding! 🎉

[arvindrajan92]

Hi @quoclam-doit

This model has not been pre-trained therefore you cannot expect the results to match that of the paper.

I have explained how I got the data here: https://github.com/arvindrajan92/DTrOCR/blob/main/examples/IAMFineTune.ipynb

Is that helpful?

[github-actions]

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.