Fantastic work! 👏 I’m curious if there are plans to open-source the training/inference code and pre-trained models for this project. If so, could you share any tentative timelines for their release? This would greatly benefit the community for reproducibility and further exploration.
Thank you for your efforts and transparency!