In contrast to a remote dataset, a local dataset is a dataset that exists in your local machine. These datasets are usually, but not always, datasets that have been pulled from V7 to your own device so you could perform some kind of specific analysis/training or so you can create a pipeline using the SDK/CLI tools we offer.
Because you can perform operations on you local datasets, it is important to notice that these can diverge from the versions you have in remotely in V7. Both SDK and the CLI offer options to update the corresponding remote datasets with the data from your local one as you can check by typing
darwin dataset help:
darwin dataset help Arguments to interact with datasets Commands: (...) create Creates a new dataset on darwin. delete-files Delete one or more files remotely. (...) import Import data to an existing (remote) dataset. local List downloaded datasets. path Print local path to dataset. pull Download a version of a dataset. push Upload data to an existing (remote) dataset. (...) remove Remove a remote or remote and local dataset.
See the SDK documentation on this topic