FedTree
Xtra-ComputingFederated learning system for gradient-boosted decision trees from Xtra Computing (MLSys 2023). It trains horizontal or vertical federated GBDTs as a standalone simulation or with separate distributed server and party binaries, on multi-core CPUs and GPUs, with homomorphic encryption, secure aggregation and differential privacy.
In this category rather than classic_ml_cv: federation is the product, and the tree learner is the payload. The repository's default branch is `cyz-grpc`, and it was last pushed 2025-01-20.
Openness
5 high confidence- license
- Apache-2.0(OSI
- source
- public(the whole system)
- core features withheld
- no — a university research system with no paid edition
FedTree is an Apache-2.0 research system with no paid tier or held-back component.
- https://raw.githubusercontent.com/Xtra-Computing/FedTree/HEAD/LICENSE recorded 2026-09-27
The LICENSE body: the verbatim Apache License, Version 2.0 text with no added clause.
- https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Xtra-Computing%2FFedTree recorded 2026-09-27
Repository record for Xtra-Computing/FedTree: license apache-2.0, description "A tree-based federated learning system (MLSys 2023)", not archived, not a fork, pushed 2025-01-20, 153 stars.
- https://ungh.cc/repos/Xtra-Computing/FedTree/files/cyz-grpc recorded 2026-09-27
The 155-file tree of Xtra-Computing/FedTree at `cyz-grpc`: one root LICENSE, and no ee/, enterprise/, commercial/, proprietary/ or premium/ path anywhere in it. A python/LICENSE also sits under the Python bindings. Path absence only, paired with the statement cited beside it.
Adoption
1 medium confidencePyPI downloads of `fedtree`, the project's Python package.
- https://pypi.org/pypi/fedtree/json recorded 2026-09-27
PyPI `fedtree` 1.0.5, uploaded 2023-01-26, "A federated learning library for trees", homepage https://github.com/Xtra-Computing/FedTree.
- https://pypistats.org/api/packages/fedtree/recent recorded 2026-09-27
last_month = 6 downloads for the package `fedtree`.
Capability
3 medium confidenceFedTree runs real distributed federations with encryption, secure aggregation and differential privacy, but only for gradient-boosted trees, and parties join without any identity check. That puts it level with Plato.
- https://raw.githubusercontent.com/Xtra-Computing/FedTree/HEAD/README.md recorded 2026-09-27
README: "Federated training of gradient boosting decision trees"; "Supporting homomorphic encryption, secure aggregation and differential privacy"; Standalone Simulation, Distributed Horizontal FedTree and Distributed Vertical FedTree via FedTree-distributed-server and FedTree-distributed-party.
Verified 2026-09-27