TESS Quiet Stars: ML ready delta flux dataset and anomaly catalog (Sectors 1–5)

Dataset accompanying the paper "Recovering Pipeline-Suppressed Stellar Rotation Signals in TESS Using Delta Flux Autoencoders" (Baig 2026). Contains the ML-ready flux matrix, anomaly catalogs, trained model weights, and supporting catalogs for 33,135 non-variable TESS stars from Sectors 1–5. Files:- flux_matrix.parquet (950 MB): 50,287 star-sector records × 1,024 time points × 3 channels (SAP, PDCSAP, delta flux)- anomaly_catalog_clean.parquet: 88-star clean anomaly catalog with periods, GAIA flags, and repeatability scores- anomaly_clusters_validated.parquet: Full 100-star validated catalog- period_catalog.parquet: Batch Lomb-Scargle periods for all 100 anomaly stars- boundary_hit_periods.parquet: Two-sector period recovery for 16 boundary-hit stars- gaia_binary_check.parquet: GAIA DR3 RUWE and binary flags- autoencoder_sap.pt: Trained SAP autoencoder weights (PyTorch)- autoencoder_delta.pt: Trained delta flux autoencoder weights (PyTorch) Code: https://github.com/AliBaig-xD/tess-quiet Paper: https://doi.org/10.5281/zenodo.19654856