Training PaddleOCR Detection and Recognition Models for Vertical Text

[RunSongham]

Hello everyone,
I'm new to PaddleOCR and machine learning in general, so please bear with me if this is a basic question.
Current situation:
I'm testing the Korean language model and noticed that:

Horizontal text detection and recognition work well
Vertical text detection and recognition don't work at all

From what I've read, this is because the Korean model wasn't trained on vertical text layouts.
What I'm trying to do:
I'd like to fine-tune the model using vertical text data. I've prepared a dataset with images sized at 520x110 pixels (sample attached).
My questions:

1. Is this approach correct for a beginner? Should I be doing something else instead?
2. Is the 520x110 resolution appropriate for training?
3. Are there any beginner-friendly guides or tips for fine-tuning on vertical text?

I'm still learning, so any guidance would be really helpful!
Thank you in advance!

[sample]

[liuhongen1234567]

Hello, you can refer to the PP-OCRv5 documentation https://www.paddleocr.ai/latest/en/version3.x/module_usage/text_recognition.html#3-quick-start for specific training, evaluation, and export details. In PaddleOCR, vertical text needs to be rotated to horizontal for training.