PredANNpp-NMEDT-SongID-FullScratch-ep3500-seed42

Model description

This repository contains a PredANN++ PyTorch Lightning checkpoint for EEG-based music representation learning and/or song identification.

• Canonical repository: Shogo-Noguchi/PredANNpp-NMEDT-SongID-FullScratch-ep3500-seed42
• Checkpoint file: predannpp_nmedt_songid_fullscratch_ep3500_seed42.ckpt
• Stage: fullscratch-songid
• Target / teacher representation: FullScratch
• Architecture: EncoderOnly classifier
• Random seed: 42
• SHA256: e46cc56003a8f00815d690a10fffc30acc3ddb5507b651e1ffc896082a80416b

This is a full-scratch baseline trained without representation pretraining. It is an encoder-only Song ID classifier for NMED-T.

The repository name includes NMEDT because the checkpoint is task-specific to the NMED-T Song ID classification setup.

Capabilities

Direct Song ID logits for 3-second EEG segments; no decoder and no masked teacher-token prediction objective.

Input and output

• Input EEG: 128 channels, 125 Hz, 3-second segments, following the PredANN++ / NMED-T preprocessing pipeline.
• Output: depends on stage. Pretraining checkpoints expose the multitask pretraining outputs; the full-scratch checkpoint outputs 10-class Song ID logits.

Training data

• Dataset: NMED-T (Naturalistic Music EEG Dataset – Tempo), 10 songs, 20 subjects, trial=1, as used in the PredANN++ experiments.
• Teacher / target source: no teacher representation; direct Song ID supervision only.

Training procedure

Encoder-only Song ID classification for 3500 epochs from random initialization, seed 42. No pretraining checkpoint is used.

Intended use

Baseline reproduction, comparison against representation-pretrained PredANN++ models, and NMED-T Song ID evaluation.

Not intended use

• Medical diagnosis, clinical decision making, or biometric identification.
• Commercial use without checking the PredANN++ code license, NMED-T terms, and upstream model/feature licenses.
• General EEG representation learning or masked-token pretraining evaluation; this checkpoint is a classifier baseline.

License and upstream dependencies

No MuQ/MusicGen teacher features are used for this baseline. The same CC-BY-NC-4.0 checkpoint license is used for collection-level compatibility and to avoid accidental over-permissive reuse of NMED-T-derived training artifacts.

Reproducibility notes

• The original source path at release time was: /data/Backup_AkamaUbuntu/home/sony_csl/workspace/noguchi/work/mind-model/Surprisal_Model/codes_3s/best_checkpoints/SEED42Full/LastCKP3500/last.ckpt.
• metadata.json stores the standardized release metadata.
• SHA256SUMS stores the checkpoint checksum.
• Use the PredANN++ GitHub repository for model definitions and evaluation scripts.

Links

• Project page: https://shogonoguchi.github.io/PredANNpp/
• GitHub: https://github.com/ShogoNoguchi/PredANNpp
• Hugging Face collection: https://huggingface.co/collections/Shogo-Noguchi/predann-models
• Paper: https://arxiv.org/abs/2603.03190

Citation

If you use this checkpoint, cite the PredANN++ paper and the NMED-T dataset. For MuQ / MusicGen-derived teacher features, also cite the relevant upstream model or toolkit.

@misc{noguchi2026expectationacousticneuralnetwork,
  title={Expectation and Acoustic Neural Network Representations Enhance Music Identification from Brain Activity},
  author={Shogo Noguchi and Taketo Akama and Tai Nakamura and Shun Minamikawa and Natalia Polouliakh},
  year={2026},
  eprint={2603.03190},
  archivePrefix={arXiv},
  primaryClass={cs.AI},
  url={https://arxiv.org/abs/2603.03190}
}