Federated learning - Data Owner

Participant in Federated Learning scheme that can take the role of training a local instance of model or calculate the error of the federated model using local data.

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

• python3
• docker
• docker-compose

Installing

A step by step series that tell you how to get a development env running

git clone git@github.com:DeltaML/federated-learning-poc.git
cd federated-learning-poc/
python3 -m venv venv
source venv/bin/activate
pip install -r data_owner/requirements.txt

Run

Using command line

    gunicorn -b "0.0.0.0:8080" --chdir data_owner/ wsgi:app --preload

Using Docker

Select LOCAL_PORT number to run some containers

    docker build -t federated-learning-data_owner --rm -f data_owner/Dockerfile .
    docker run --rm -it -p <LOCAL_PORT>:5000 federated-learning-data_owner

docker run --rm -it -p 5001:5000 4001:4000 federated-learning-data_owner

Using Pycharm

Script Path: .../data_owner/virtualenv/bin/gunicorn
Parameters: -b "0.0.0.0:5000" wsgi:app --preload
Working directory: ../data_owner

Usage

Upload file

 curl -v  -F ‘data=/data_owner/dataset/file_1.csv’ "http://localhost:5000/weights"
 curl -v -X POST file=@"data_owner/dataset/file_1.csv" http://localhost:5000/datasets

Process weights

curl -v -H "Content-Type: application/json" -X POST -d '{"type": "LINEAR_REGRESSION", "public_key": "XXXXXXXXXXXXXXXX"}' "http://localhost:5000/weights"

Gradient step

curl -v -H "Content-Type: application/json" -X POST -d '{"gradient":[{"w1":1}]}'  "http://localhost:9090/step"

Get Model from data owner

curl -v -H "Content-Type: application/json" -X GET "http://localhost:9090/model"

Data Owner configuration

N_SEGMENTS = 5
FEDERATED_AGGREGATOR_HOST = "http://cte_federated_aggregator:8080"
ETA = 1.5
REGISTRATION_ENABLE = True
ACTIVE_ENCRYPTION = False

Configuration details

• NSEGMENTS: _TODO
• FEDERATEDAGGREGATORHOST: __TODO
• ETA: TODO
• REGISTRATIONENABLE: _TODO
• ACTIVEENCRYPTION: _TODO

Versioning

We use SemVer for versioning. For the versions available, see the tags on this repository.

Authors

• Fabrizio Graffe - Dev - GFibrizo
• Agustin Rojas - Dev - agrojas

See also the list of contributors who participated in this project.

License

This project is licensed under the MIT License - see the LICENSE.md file for details