We are pleased to announce that the first version of LoVA’s training and inference code is now available on our GitHub repository! The corresponding model weights have also been released, and we invite you to click the url, explore the code and its implementation.

Please note that the current version is relatively simple and may require further refinements due to time constraints during development. We plan to release a more polished, optimized, and well-documented version following its acceptance.

Your feedback and suggestions are greatly appreciated as we continue to improve LoVA. Thank you for your support!