Migrating Datasets to the Tensor Database

Migrating datasets to the Tensor Database

How to migrate existing Deep Lake datasets to the Tensor Database

Datasets are created in the Tensor Database by specifying the dataset_path = "hub://<org_id>/<dataset_name>" and runtime = {"db_engine": True}) during dataset creation. If datasets are currently stored locally, in your could, or in non-database Activeloop storage, they can be migrated to the Tensor Database using:

ds_tensor_db = deeplake.deepcopy(src = <current_path>, 
                                 dest = "hub://<org_id>/<dataset_name>", 
                                 runtime = {"db_engine": True}, 
                                 src_creds = {<creds_dict>}, # Only necessary if src is in your cloud
                                 )