Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

If I want to train the network with another dataset, what should I do? #2

Open
whereitogo opened this issue Dec 11, 2019 · 2 comments

Comments

@whereitogo
Copy link

No description provided.

@raidastauras
Copy link
Owner

raidastauras commented Dec 13, 2019

Unfortunately, there's no easy answer to your question.

This project is pretty complicated, especially the data preparation and pre-processing part. For data preparation I borrowed most of the code from other popular repos, e.g. Pay20Y. Data prep code is stored in synthtext.py and icdar.py. Unfortunately it is not easily readable and digestible. It took me a couple of full days to read into it and modify it according to my needs.

As of now I'm not aware of any good example of well written and readable code for data prep for FOTS and sadly, this repo is not an exception (data prep code should ideally be totally refactored).

To go back to your question, my advise would be to carefully investigate how does the model training input look like and how is it used. Then move on from there.

@whereitogo
Copy link
Author

Thank you!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants