Quick Start

Training

Edit experiment name in train.run:

sbatch train.run

Generate mega-array for all scenarios:

python generate_inpaint_jobs.py \
  -e "paper-2025-06" \
  --scenarios "0-31" \
  --start_date "2022-10-12" \
  --end_date "2023-05-15"

This creates: - Job list: inpaint_jobs_paper-2025-06_all_scenarios.txt - SLURM script: inpaint_array_paper-2025-06_all_scenarios.run

Submit all scenarios:

sbatch inpaint_array_paper-2025-06_all_scenarios.run

Aggregate results (run locally):

python aggregate_results.py \
  -e "paper-2025-06_inpainting" \
  --compute_wis \
  --create_ensemble \
  --plot_results