Training Dataset and Trained Model Weights for "Accelerating Chemical Kinetics for Exoplanet Atmospheres using Neural Networks"

Trained model weights and dataset-provenance metadata accompanying the paper "Accelerating Chemical Kinetics for Exoplanet Atmospheres using Neural Networks" (Malsky et al., submitted to The Astrophysical Journal). Includes the published PyTorch Lightning checkpoint (best.ckpt), the exported deployable model (physical_model_k1_cpu.pt2), training metrics/logs, and the JSON records describing how the training/validation/test dataset was constructed and normalized (normalization.json, shard_index.json, preprocess_report.json, preprocessing_summary.json, dataset_config_snapshot.json). The emulator code is available at https://github.com/imalsky/Chemulator and the Robertson benchmark code at https://github.com/imalsky/robertson-emulator.