Firstly, I want to express my sincere gratitude for your outstanding work on this project. Your efforts have been incredibly inspiring and have helped me a lot in my own work.

I noticed that the project already includes training code and some trained models, which is very helpful. However, I was wondering if it’s possible to open-source the complete training model code and configurations? This would allow users to dive deeper into the training process and make further optimizations based on their needs.