You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, benchmark datasets are considered stable and immutable over time. However, in the future we might want to publish new versions of dataset, i.e., add more events/traces or further features. For this, SeisBench should be able to provide different versions of datasets.
This issues is mostly here as a reminder and is only be intended to be addressed, once actually a new version of a dataset is supposed to be published.
The text was updated successfully, but these errors were encountered:
The only add-on comment to this would be - from an implementation perspective - if the old versions of datasets are to be kept available to allow for backward compatibility with users implemented workflows, we may also just need to check what the storage limits are on dCache, and whether we can extend them.
Good point! I think it's very relevant to keep old versions available to ensure reproducibility. For now, I think we shouldn't have any issues with dCache, as compared to their particle physics data, the size of our data is negligible.
Currently, benchmark datasets are considered stable and immutable over time. However, in the future we might want to publish new versions of dataset, i.e., add more events/traces or further features. For this, SeisBench should be able to provide different versions of datasets.
This issues is mostly here as a reminder and is only be intended to be addressed, once actually a new version of a dataset is supposed to be published.
The text was updated successfully, but these errors were encountered: