How anurag2796/fedlearn-platform works
FedLearn is a federated-learning product spanning a web dashboard, Spring API, Python FL framework/runtime, deployable clients, desktop tooling, and a native mobile client. The primary workflow is project/model setup, backend-managed server launch, client training over gRPC, progress streaming to the dashboard, and persisted results. The README documents additional recipes, strategies, simulation, benchmarking, marketplace and inference capabilities. Some runtime paths are documented but not sampled, so those links are dashed and unverified.
Main entry points: App.tsx Web Dashboard · InferenceController.java REST API · server.py FL Coordinator · client.py FL Runtime
Components 19
Web Product
- Web Dashboard
frontend/src/App.tsx - Project Setup
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/service/ProjectService.java - Benchmarking
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/controller/BenchmarkIngestController.java - Model Marketplace
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/controller/MarketplaceController.java
Platform Backend
- REST API
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/controller/InferenceController.java - Access Control
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/service/AuthorizationService.java - Server Lifecycle
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/orchestration/FlServerManager.java - Platform State
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/model/Project.java - Live Updates
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/service/WebSocketService.java - Model Inference
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/service/InferenceService.java
Federated Learning
- FL Runtime
fl-runtime/client.py - Recipe Catalog
fl-runtime/recipes.py - FL Coordinator
framework/src/fedlearn/server/server.py - gRPC Service
framework/src/fedlearn/server/grpc_servicer.py
Client Apps
- Training Client
framework/src/fedlearn/client/grpc_client.py - Desktop App
fedlearn-desktop/src/renderer/App.tsx - Mobile Client
mobile_client/src/App.tsx
Other components
- Dashboard User
- PostgreSQL
Connections 21
- Dashboard User → Web Dashboard: uses
- Web Dashboard → REST API: calls
- Web Dashboard → Project Setup: configures
- Project Setup → Server Lifecycle: starts server
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/service/ProjectService.java - Server Lifecycle → FL Runtime: spawns scripts
- FL Runtime → FL Coordinator: runs framework
- FL Coordinator → gRPC Service: registers service
framework/src/fedlearn/server/server.py - Training Client → gRPC Service: connects
- Mobile Client → REST API: uses REST
- Desktop App → REST API: calls inference
- REST API → Model Inference: dispatches
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/controller/InferenceController.java - Model Inference → FL Runtime: runs scripts
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/service/InferenceService.java - Model Inference → Live Updates: streams tokens
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/service/InferenceService.java - REST API → Access Control: enforces access
- Project Setup → Platform State: persists projects
backend/fl-platform-api/src/main/java/com/federated/fl_platform_api/service/ProjectService.java - Platform State → PostgreSQL: stored in
- FL Coordinator → Live Updates: streams logs
- Recipe Catalog → FL Runtime: provides data
fl-runtime/client.py - Web Dashboard → Benchmarking: views metrics
- Web Dashboard → Model Marketplace: browses models
- FL Coordinator → Platform State: stores results