Yinpei

/

robomme_data_h5

like 2

I noticed that the repo https://huggingface.co/Yinpei/robomme_data_h5 is hosted as a model repository on HuggingFace, but its content is actually training data (HDF5 demonstration files). This causes issues when downloading with huggingface_hub:

from huggingface_hub import snapshot_download

This fails with 404 because the repo is not actually a dataset

snapshot_download(repo_id="Yinpei/robomme_data_h5", repo_type="dataset")

Have to use repo_type="model" to download successfully

snapshot_download(repo_id="Yinpei/robomme_data_h5", repo_type="model")

Would it be possible to migrate the repo to the datasets namespace (i.e., https://huggingface.co/datasets/Yinpei/robomme_data_h5)? This would be more consistent with HuggingFace conventions and avoid confusion for users trying to download the data.

Thanks!