Metadata-Version: 2.4
Name: syft-flwr
Version: 0.2.2
Summary: syft_flwr is an open source framework that facilitate federated learning projects using Flower over the SyftBox protocol
License-File: LICENSE
Requires-Python: >=3.10
Requires-Dist: flwr-datasets[vision]>=0.5.0
Requires-Dist: flwr[simulation]==1.21.0
Requires-Dist: loguru>=0.7.3
Requires-Dist: safetensors>=0.6.2
Requires-Dist: syft-rds>=0.2.2
Requires-Dist: tomli-w>=1.2.0
Requires-Dist: tomli>=2.2.1
Requires-Dist: typing-extensions>=4.13.0
Description-Content-Type: text/markdown

# syft_flwr

`syft_flwr` is an open source framework that facilitate federated learning (FL) projects using [Flower](https://github.com/adap/flower) over the [SyftBox](https://github.com/OpenMined/syftbox) protocol

![FL Training Process](notebooks/fl-diabetes-prediction/images/fltraining.gif)

## Example Usages
Please look at the `notebooks/` folder for example use cases:
-  [FL diabetes prediction](notebooks/fl-diabetes-prediction/README.md) shows how to train a federated model over distributed machines for multiple rounds
-  [Federated analytics](notebooks/federated-analytics-diabetes/README.md) shows how to query statistics from private datasets from distributed machines and then aggregate them
-  [FedRAG (Federated RAG)](notebooks/fedrag/README.md) demonstrates privacy-preserving question answering using Retrieval Augmented Generation across distributed document sources with remote data science workflow