Skip to content

Latest commit

 

History

4 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 

Repository files navigation

DeepLynx-Run

DeepLynx-Run is an orchestration platform for running DeepLynx API driven workflows as Apache Airflow DAGs. It provides the infrastructure to schedule, monitor, and manage tasks that read from and write to DeepLynx, with federated learning as one of several supported workload types.

federated_learning

One of DeepLynx-Run's supported workloads is a production oriented federated learning framework built on the Flower framework, supporting PyTorch and XGBoost. It provides a broad set of aggregation strategies (FedAvg, FedProx, FedAdagrad, FedYogi, Krum, Bulyan, and variants), differential privacy with configurable clipping and privacy budget tracking, secure aggregation, and fault tolerant coordination for unreliable clients. It includes Docker based deployment and Kubernetes orchestration with preflight security checks (RBAC, pod security, network policy), integrated into DeepLynx-Run's Airflow pipelines.

Keywords

  • federated learning
  • Flower framework
  • PyTorch
  • XGBoost
  • differential privacy
  • secure aggregation
  • fault tolerance
  • distributed machine learning
  • privacy preserving machine learning
  • Kubernetes
  • Apache Airflow
  • DeepLynx
  • model aggregation
  • privacy budget
  • edge computing

About

DeepLynx-Run framework for semiconductor manufacturing that demonstrates collaborative machine learning on distributed local datasets without sharing proprietary data. Enables privacy-preserving model training across organizations for applications such as process optimization, defect detection, yield forecasting, and equipment failure prediction.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Used by

Contributors

Languages