HuBERT and data2vec retrained on music

[lcc-404]

In original paper , the speech SSL models HuBERT and data2vec are selected as baselines and retrain them on music data, will these two checkpoints be released?

[Laubeee]

see https://huggingface.co/m-a-p/MERT-v0
the data2vec model is there as music2vec. I don't think they re-trained the original HuBERT though?