GraphCanon updated 2w · GitHub synced 2w
Decision brief
Harmonia supports federated learning with differential privacy modules and GitOps-inspired architecture, designed for both research and production usage.
Good fit when
- When needing frameworks that incorporate differential privacy directly into federated learning processes
- For environments preferring a container-based plug-in architecture to integrate existing ML workflows
Avoid when
- If GitOps-inspired workflows are not aligned with your team's operational practices
- In scenarios where the use of Go is less preferred among development teams
Observed Jul 17, 2026 · Source: enrich:decision_facts
Verify the decision
Maintenance and security
Full trust report- Maintenance
- Dormant (2143d since push)
- As of 2w
- Provenance
- Not a fork · Organization account
- As of 2w
- Security (OSV)
- No lockfile
- As of 1mo
Public GitHub metadata and optional OSV scans. Signals, not a guarantee. Trust methodology.
Install
go get github.com/ailabstw/harmoniaSimilar tools
Same-category neighbours. No typed graph edges are catalogued for this tool yet.
Evidence and technical details
Sourced facts, taxonomy, compatibility claims, README excerpt, and machine-readable endpoints.
Overview
Harmonia is an open-source project that includes systems/infrastructures and libraries to facilitate the adoption of federated learning for researches and production usage. It supports differential privacy modules and GitOps-inspired architecture.
Capability facts
- Languages
- go
Source: github.language · Aug 4, 2026
Categories
Tags
README
Harmonia
Harmonia is an open source project aiming at developing systems/infrastructures and libraries to ease the adoption of federated learning (abbreviated to FL) for researches and production usage. It is named Harmonia, the Greek goddess of harmony, to reflect the spirit of federated learning; that is, multiple parities collaboratively build a ML model for the common good.
The first release includes Harmonia-operator SDK and differential privacy modules (https://github.com/ailabstw/blurnn). We welcome contributions of new aggregation algorithms, privacy mechanism, datasets, etc. Let's work together to flourish the growth of federated learning.
FL System Architecture
The design of the Harmonia system is inspired by GitOps. GitOps is centerred around a git repository, which maintains the desired states in the production environment. An automated process makes the production environment match the described state in the repository. In Harmonia, training plans (or simply FL parameters), global models, and local models are kept in git repositories. Updates to these repositories trigger FL system state transitions. These automates the FL training processes. A participant in a federated training is composed of an Operator container and an Application container. An Operator container is in charge of maintaining the FL system states, and communicates with an Application container via gRPC. Local training and aggerator applications are encapsulated in Application containers. This container based architecture enables quick plug-in of existing ML workflows.
Documentation
- docs
- docs/get-started: A step-by-step example tutorial
- docs/sdk: Detailed sdk document
Build
To build the Harmonia Operator,
$ make all
for harmonia/operator image
Example
See examples/mnist
Get Started
See docs/get-started
For agents
This page has a .md twin and JSON over the API.