Federated Learning Simulation with Flower & PyTorch
Budget: $20 – $50 USD
I'm seeking an expert to develop a federated learning simulation using the Flower framework and PyTorch. The simulation should implement all major FL algorithms (FedAvg, FedProx, FedAdam, FedYogi, SCAFFOLD, COOP) on the MNIST, CIFAR-10/100, FEMNIST, and Shakespeare datasets.
Requirements:
- Fully automated and configurable via pyproject.toml
- Test on IID and non-IID scenarios
- Collect key metrics: fairness, robustness, privacy, communication, heterogeneity
- Output thousands of data points + structured plots for analysis
- Save results as CSV, JSON, plots, and logs in organized folders
- Deliverables: working system, sample results, README
Ideal Skills & Experience:
- Proficient in Flower framework and PyTorch
- Strong background in federated learning
- Experience with data handling and automation
- Familiarity with the specified datasets and evaluation metrics
Requirements:
- Fully automated and configurable via pyproject.toml
- Test on IID and non-IID scenarios
- Collect key metrics: fairness, robustness, privacy, communication, heterogeneity
- Output thousands of data points + structured plots for analysis
- Save results as CSV, JSON, plots, and logs in organized folders
- Deliverables: working system, sample results, README
Ideal Skills & Experience:
- Proficient in Flower framework and PyTorch
- Strong background in federated learning
- Experience with data handling and automation
- Familiarity with the specified datasets and evaluation metrics