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

fine-tuning on a custom dataset with the official craft-pytorch weights #81

Open
alikaz3mi opened this issue Aug 9, 2021 · 4 comments

Comments

@alikaz3mi
Copy link

Hello, is it possible to fine-tune your model on a custom dataset with Craft-PyTorch released weights too?

@gl94
Copy link

gl94 commented Sep 3, 2021

Yes, I just made it using the IC15 traing code. I labeled my dataset the same format as IC15.

@tahakoudri
Copy link

Hello, gl94. Can you explain how did you manage do to it and if it works well ?

@gl94
Copy link

gl94 commented May 19, 2022

Hello, gl94. Can you explain how did you manage do to it and if it works well ?

It's kinda a little bit tricky, you need to use the old branch of this repo, and change some code in the trainic15data.py and maybe some function in the dataloader to get a better model. After all this been done you'll find that craft is really hard to train. I rather recommend you using dbnet in the PPocr. It's easy to train and performs very well.

@SkygirlLuna
Copy link

@gl94 What is PPocr?

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

4 participants