Releases: spdlearn/spd_learn
Releases · spdlearn/spd_learn
v0.1
SPD Learn v0.1.0 Release Notes:
- 7 Neural Network Models: SPDNet, EEGSPDNet, TSMNet, TensorCSPNet, PhaseSPDNet, GREEN, MAtt
- 10+ SPD Layers: BiMap, ReEig, LogEig, ExpEig, SqrtEig, InvSqrtEig, PowerEig, VecMat, CovLayer, Shrinkage
- 5 Batch Normalization variants: SPDBatchNorm, TSMBatchNorm, AdaMomSPDBatchNorm, DomainSPDBatchNorm, TrackingMeanBatchNorm
- 4 Riemannian Metrics: Affine Invariant Riemannian (AIRM), LogEuclidean, LogCholesky, BuresWasserstein
- Functional Operations: Matrix operations, geodesics, statistics, parallel transport, covariance estimation
- Documentation: Installation guide, user guide, theory section, API reference, examples gallery
- Technical specs: Python 3.11-3.13, PyTorch 2.0+, BSD-3-Clause license
- Acknowledgments: Inria, CNRS, CEA, Université Paris-Saclay, ATR, USMB, UCSD
- Citation: BibTeX entry included
The file follows the https://keepachangelog.com/ format and is ready for your v0.1.0 release.