Compare DAC, TFM, and LaBraM EEG tokens using shared classifiers: cached EEG + fixed splits → tokenization → classifier training → results.
| Directory | Contents |
|---|---|
| dataset cache/ | EEG tensors, NeuroAI splits, metadata, and export checks. |
| tokenization/ | Pretrained adapters, in-memory inference, and optional token export. |
| token_cache/ | Generated tokens, codebooks, and provenance. |
| models/ | Transformer and bag-of-tokens classifiers. |
| training/ | Shared classifier runner for cached or live tokens, fixed splits, and comparison results. |
| synthetic_data/ | Controlled EEG datasets and classifier experiments. |
| results/ | Experiment runners, metrics, plots, and saved configurations. |
| tests/ | Dataset, tokenization, classifier, and experiment checks. |