Skip to content

User Guide

The user guide covers how to run pamica in practice and how its results relate to the reference implementation.

  • Backends & Devices: the available compute backends (PyTorch natural-gradient EM, optional MLX, legacy NumPy), how to select one through AMICA and AMICAICA, device selection (CUDA / CPU / MPS), float32 vs float64, and performance guidance on real EEG.
  • EEGLAB interoperability: writing a fit as an EEGLAB amicaout directory and reading it with loadmodout15.
  • Validation & Parity: how correctness is defined as parity with the Fortran reference, the validation harness, and how cross-backend equivalence depends on data adequacy.
  • Differences vs AMICA: every place pamica deliberately behaves differently from the Fortran reference, why, and how to restore the reference behavior.