New Release! Git-backed Machine Learning Model Registry for all your model management needs.
Another powerful feature of MLEM is linking. Links are lightweight objects that represent MLEM Objects in other locations. You can reference links anywhere you need to specify MLEM Object bot in API and CLI.
Since links are also a type of MLEM Object, they share the same internal logic.
For example, they are saved under the .mlem/link directory. To load an
instance of MlemLink (and not the object it references) provide
follow_links=False to load_meta method.
The content of the link is very lightweight and consists of the following fields:
link_type - type of referenced objectfs) as in
hereobject_type="link""Links can be created via mlem link command or mlem.api.link() API, as well
as the MlemObject.make_link() method.
You can create relative links inside the same repository, which will basically create an alias for that object.
Also, since links can target specific commits, tags or branches in a versioned repository, they can be used in a variety of different scenarios, for example to create a centralized Model Registry.