Baku Metro FedAvg GRU

Summary

Federated GRU baseline for station-level next-day passenger entry forecasting under the one-station-one-client setting.

Task

Station-level next-day passenger entry forecasting.

Training Data

Processed dataset reference: olaflaitinen/baku-metro-station-flow

Architecture

Compact GRU regressor trained with Flower FedAvg aggregation under the one-station-one-client setting.

Training Procedure

Clients train local GRU updates on station-specific sequences. The server aggregates model parameters with FedAvg across deterministic training rounds.

Evaluation

Metrics are exported from the local artifact bundle and include MAE and RMSE on held-out dates.

Intended Use

  • reproducible federated learning baselines
  • transportation demand forecasting experiments

Out-of-Scope Use

  • live passenger operations
  • identity analytics
  • route inference

Ethical and Operational Considerations

This model operates on aggregated station-level entry counts and should not be interpreted as a claim about individual passengers or operational causality.

Reproducibility

See the repository configuration files, requirements lock, and artifact manifest for exact experimental settings.

License

The repository code, documentation source, configuration, and release metadata are licensed under Apache-2.0. The trained model artifact in this repository is released under the same project license unless external platform terms impose additional distribution mechanics.

Citation

Please cite the repository software record in CITATION.cff.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 1 Ask for provider support

Dataset used to train olaflaitinen/baku-metro-fedavg-gru