These tutorials cover a range of frameworks, models, and datasets to help you get started with OpenFL. This directory provides notebooks for the Workflow API, one of two ways to run Federated Learning experiments in OpenFL.

Install OpenFL following the installation guide and activate workflow API.

fx experimental activate

• MNIST: An introductory guide to Workflow Interface for federated learning with a small PyTorch CNN model on the MNIST dataset.
• Aggregator Validation: Shows how to perform validation on the aggregator after training using OpenFL Workflow Interface.
• Cyclic Institutional Incremental Learning: Demonstrates cyclic institutional incremental learning using OpenFL Workflow Interface.
• Keras MNIST with CPU: Trains a CNN on MNIST using Keras on CPU with OpenFL Workflow Interface.
• Keras MNIST with GPU: Uses Keras to train a CNN on MNIST with GPU support via OpenFL Workflow Interface.
• MNIST XPU: Trains a CNN on MNIST using Intel Data Center GPU Max Series with OpenFL Workflow Interface.
• Numpy Linear Regression: Implements linear regression with MSE loss using Numpy and OpenFL Workflow Interface.
• Scikit Learn Linear Regression: Trains a linear regression model with Ridge regularization using scikit-learn and OpenFL Workflow Interface.
• Exclusive GPUs with Ray: Utilizes Ray Backend for exclusive GPU allocation in federated learning with OpenFL.
• MNIST Watermarking: Demonstrates embedding watermark in a DL model trained on MNIST in a federated learning setup.
• FedProx with Synthetic non-IID: Compares FedProx and FedAvg algorithms on a synthetic non-IID dataset using OpenFL Workflow Interface.
• MNIST Aggregator Validation Ray Watermarking: Combines aggregator validation, watermarking, and multi-GPU training using Ray Backend in OpenFL.
• Federated FedProx PyTorch MNIST Workflow Tutorial: Implements FedProx algorithm for distributed training on MNIST using PyTorch and Workflow API.
• Keras MNIST with FedProx: Trains a TensorFlow Keras model on MNIST using OpenFL Workflow Interface with FedProx optimization.
• Federated Evaluation MNIST Workflow Tutorial: Shows how to implement Federated Evaluation using Workflow API.

Refer to the respective README files for detailed information on the executing these tutorials.

• CrowdGuard: Implementing CrowdGuard to mitigate backdoor attacks in Federated Learning.
• Federated Runtime: Demonstrates use of FederatedRuntime for taking workflow API from simulation to deployment.
• Differential Privacy: Differential Privacy in Federated Learning.
• LLM Fine-tuning: Fine-tuning a Large Language Model (LLM) using OpenFL Workflow Interface.
• Privacy Meter: Demonstrates integration and use of ML Privacy Meter library with OpenFL to audit privacy loss in federated learning models.