Registered users can ask their own questions, contribute to discussions, and be part of the Community!
Is it possible to use a custom keras model (hdf5-format), trained outside dataiku, in a flow.
A bit like:
Fyi:- Given that dataiku is deployed on AWS (thus on efs) where I can't use folder.get_path() .
For non-filesystem managed folders, you can use instead folder.get_download_stream (See https://doc.dataiku.com/dss/latest/python-api/managed_folders.html#dataiku.core.managed_folder.Folde... for more details) to retrieve your model.
Hope this helps,