Model Registry

The Model Registry components at frontend/ml-workbench/src/components/MLflow/ and frontend/ml-workbench/src/components/ModelRegistry/ provide MLflow-integrated model cataloging with versioning, staging, and deployment tracking.

Components

ModelRegistry (MLflow Integration)

File: frontend/ml-workbench/src/components/MLflow/ModelRegistry.tsx

Displays registered models from MLflow with version history, stage transitions (Staging, Production, Archived), and metadata.

ModelRegistry (Standalone)

File: frontend/ml-workbench/src/components/ModelRegistry/index.tsx

Platform-native model catalog with search, filtering, and model comparison.

Types

// frontend/ml-workbench/src/components/MLflow/types.ts
interface RegisteredModel {
  name: string;
  creationTimestamp: number;
  lastUpdatedTimestamp: number;
  description: string;
  latestVersions: ModelVersion[];
  tags: Record<string, string>;
}
 
interface ModelVersion {
  name: string;
  version: string;
  creationTimestamp: number;
  currentStage: 'None' | 'Staging' | 'Production' | 'Archived';
  source: string;
  runId: string;
  status: 'READY' | 'PENDING_REGISTRATION' | 'FAILED_REGISTRATION';
}

Advanced Features

ModelVersioning

File: frontend/ml-workbench/src/components/advanced/ModelVersioning.tsx

Visual version comparison with metric diffs, artifact inspection, and lineage tracking across model versions.