zestimatr

Spectroscopic redshift estimation with uncertainty quantification.

zestimatr uses a 1D convolutional neural network to predict redshifts from high-resolution spectra, providing both point estimates and calibrated uncertainties.

API Reference

• Inference
  • InferenceDataset
  • download_pretrained()
  • load_model()
  • predict_redshifts()
• Model
  • ResConvBlock
  • ZHead1D
• Dataset
  • SpectraForZHead
• Metrics
  • compute_calibration_metrics()
  • compute_metrics()
• Emission Lines
  • detect_emission_lines()
  • plot_spectrum()
• Plotting
  • plot_predictions()