wallaroo.model_registry
An instance of an external Model Registry. Currently, the supported registries are: DataBricks
Base constructor.
Each object requires:
- a GraphQL client - in order to fill its missing members dynamically
- an initial data blob - typically from unserialized JSON, contains at
- least the data for required members (typically the object's primary key) and optionally other data members.
Parameters
- name str A descriptive name to set
- path str The DBFS path to the artifact file to upload.
- framework Framework The ML framework that this model uses (ex. Framework.ONNX).
- requirements List[str] An optional list of python requirements needed to run this model.
- input_schema pa.Schema A PyArrow schema describing inputs to this model.
- output_schema pa.Schema A PyArrow schema describing the outputs of this model.
Returns
Model An instance of the Wallaroo model that is being created.
Wraps a list of Model Registries for display in a display-aware environment like Jupyter.
Inherited Members
- builtins.list
- list
- clear
- copy
- append
- insert
- extend
- pop
- remove
- index
- count
- reverse
- sort
An MLFlow version of a [RegisteredModel]. Versions start at 1.
Attributes: creation_timestamp (int): Timestamp in milliseconds from epoch current_stage (str): last_updated_timestamp (int): Timestamp in milliseconds from epoch name (str): run_id (str): source (str): status (str): version (str): description (Union[None, Unset, str]): run_link (Union[None, Unset, str]): user_id (Union[None, Unset, str]):
Method generated by attrs for class RegisteredModelVersion.
def upload(self, framework: Framework, requirements: List[str] = []): # Uploads this Model's source code to your Wallaroo instance.
#:param: framework Framework The ML framework that this model uses (ex. Framework.ONNX).
#:param: requirements List[str] An optional list of python requirements needed to run this model.
#:return: Model An instance of the Wallaroo model that is being created.
#
from .wallaroo_ml_ops_api_client.api.model.upload_from_registry import sync
from .wallaroo_ml_ops_api_client.models.upload_from_registry_request import (
UploadFromRegistryRequest,
)
from .wallaroo_ml_ops_api_client.models.conversion import Conversion
ret = sync(
client=self._client.mlops(),
body=UploadFromRegistryRequest(
Conversion(
framework=framework._to_openapi_framework(),
requirements=requirements,
),
self.name,
self.registry_id,
self.version,
"private",
self._client.get_current_workspace()._id,
),
)
if ret is None:
raise Exception("Failed to start upload.")
return Model(client=self._client, data={"id": ret.model_id})
Inherited Members
- wallaroo.wallaroo_ml_ops_api_client.models.registered_model_version.RegisteredModelVersion
- creation_timestamp
- current_stage
- last_updated_timestamp
- name
- run_id
- source
- status
- version
- description
- run_link
- user_id
- additional_properties
- to_dict
- from_dict
- additional_keys
Built-in mutable sequence.
If no argument is given, the constructor creates a new empty list. The argument must be an iterable if specified.
Inherited Members
- builtins.list
- list
- clear
- copy
- append
- insert
- extend
- pop
- remove
- index
- count
- reverse
- sort
Wraps a list of Model Registries for display in a display-aware environment like Jupyter.
Inherited Members
- builtins.list
- list
- clear
- copy
- append
- insert
- extend
- pop
- remove
- index
- count
- reverse
- sort
This is a reference to a Registered Model in an external registry.
This is a list of Registered Models in an external registry.
Inherited Members
- builtins.list
- list
- clear
- copy
- append
- insert
- extend
- pop
- remove
- index
- count
- reverse
- sort