Access the federated modeling console
In the federated modeling console, you can use offline data from nodes as modeling samples to create federated models and then distribute them to each node.
Prerequisites
Before accessing the federated modeling console, confirm the following:
Federated nodes have been deployed and data has been loaded into them.
You have an existing federated modeling project in the organization management console or an invitation to join one.
The federated modeling project contains authorized nodes and data.
Access the federated modeling console
Log on to the Ant Privacy-Preserving Computation Platform.
In the navigation pane on the left, click Privacy Computing Projects.
On the Privacy Computing Projects page, click Created Projects or Invited Projects as needed.
Click Enter Project next to the desired Federated Learning (FL) project to open the project's federated modeling console.
Script Files: Manage your script files and folders. You can create, delete, and rename them.
Modeling Data: Contains the data available for federated modeling samples in the project. This data is sourced from the offline data that each node has authorized and is configured in the organization management console.
Script Editor: Define models using Python, with raw samples or federated tables from the project as script inputs. After a script runs, its output data becomes a new federated table that can be used by other scripts. Any models or pre-processing rules generated by the script are also saved and managed in the project.
Task Management: Tracks the execution status, logs, and results for each script task.
Model Management: Centrally manages and maintains the federated models and pre-processing rules generated by script tasks.