Hi GrittyHawk31 , can you elaborate on what you mean by metadata? Regarding models you can achieve this by defining the following in Task.init(output_uri="<S3_BUCKET>")

like some details about attributes, dataset size, formats.

Can you elaborate on how exactly you'd be saving this data?

here when we define output_uri in task_init in which format the model would be saved?

It depends on the framework I guess 🙂

like some details about attributes, dataset size, formats. here when we define output_uri in task_init in which format the model would be saved? CostlyOstrich36

Basically the model versions, training results, the things that are saved in mongodb, can we get those data to a s3 bucket in any format? CostlyOstrich36