Views
No views yet
2026_W26. On Hugging Face this snapshot should
be tagged exactly 2026_W26.models/channel/: channel-level spatial encoder and probe-confidence model.models/unit/: unit-level multimodal autoencoder + PT-GMM artifacts when available.context/: frozen PCA context volumes used by the channel-level model.split.json: authoritative train/validation/test insertion PID strings.features.json: authoritative ordered channel feature list.preprocessing/channel_stats.npz: frozen clipping and normalization statistics.config.json: model, preprocessing, data, and training configuration.metadata.json: code/environment provenance and component availability.results/: release evaluation summaries.