jsantillana/ca-mhd-weights

CA-MHD — Conditional Calibration Contracts under Coupled Load

Model description

Per-cell calibrated networks for the CA-MHD campaign. Each .pt file contains a trained PyTorch state dict for a specific load cell g and seed s. These weights reproduce Tables/Figures 1-4 of the ACM SIGCONF paper.

Paper: Juan S. Santillana. CA-MHD — Conditional Calibration Contracts under Coupled Load. ACM preprint.

GitHub: https://github.com/vectrayx/ca-mhd

Contents

• cell_g{0,0.25,0.5,1}_s{0,1,2}.pt — 12 cells total
• Training seeds: train = seed, val = 100+seed, test = 200+seed, rollout = 300+seed

Reproduce

git clone https://github.com/vectrayx/ca-mhd
cd ca-mhd
python -m venv .venv && source .venv/bin/activate
pip install -e .
python scripts/make_report.py

License

Paper: CC-BY-4.0 Code/weights: Apache-2.0