** edit - i just found JoblibSerializedFile -that should do the trick Hey Flyters, I have a workflow that should return a dataclass with the following fields: • model(sklearn/tf/pt) • dataset (another dataclass that is json serializable) • other artifacts from model training How can i turn an arbitrary type like an sklearn model into a dataclass field? I'm looking for a solution that is similar to structured datasets for dataframes and numpys. Should I use pickled files?

you can just put Flytefile or any object and then it will pickle