The PureML model registry provides a centralized location for users to store their models and manage their lifecycle collaboratively. This makes it easier for stakeholders to manage models and promotes transparency in accessing models for tests, deployment, audit, and other purposes.
With PureML, you can manage your models in the following ways:
- Create models to track and assess their relevance using generic model data.
- Automatically version models to efficiently organize model runs. PureML logs model parameters, metrics, and other metadata that changes for each registered model.
- Fetch stored models and their metadata through our API.
By using the PureML model registry, you can streamline your model management processes and facilitate collaboration among your team members. This allows for greater efficiency, transparency, and accuracy in managing your models.
Creating Models and Branches
With the PureML model module, you can perform a variety of actions related to creating and managing models and branches. Here’s an overview of the available methods:
Creating a Model To create a new model, import the pureml module and use the
import pureml pureml.model.init(label='FirstModel:dev', readme='ReadME.md')
The name of the model and the branch to be created are required parameters. You can also provide an optional readme file path.
label parameter consists of model name, branch in the following format:
For initializing a model, version is not required. So, we use <name>:<branch> as the label.
label should not contain any spaces. Special characters other than ”-” and ”_” are not allowed.
To list all available models, use the
import pureml pureml.model.list()
Creating a Branch
To create a new branch for a model, use the
import pureml pureml.model.init_branch(label='FirstModel:dev_2')
The branch name and the name of the model in which the branch will be created are required parameters.
To list all available branches for a model, use the
import pureml pureml.model.branch_list(model_name='FirstModel')
label parameter consists model name, branch in the following format,
For getting a list of branches of a model, branch and version are not required. So, we use <name> as the label.
These methods make it easy to create and manage the models and branches in PureML. By using them, you can streamline your model management workflows and improve collaboration among team members.