Could you make new weights of arxiv V2 version public?

[b1a0]

Hi,
I recently came across the V2 Erratum on Arxiv of the multiGradICON paper. There it says that you retrained the models with the fixed multimodal training strategy, since the previous one was using the same modality for model input and loss calculation. Two questions:

1. Does this concern the pre-trained weights that you made available for multigradicon?
2. If yes, could you provide the new best weights for multiGradICON, additionally to the previous version? It would be great to have them available, including information on the datasets they were trained on and which version of the code (commit maybe?) they were trained with.

Thank you very much!