MVPFormer-S backbone

Generatively pre-trained MVPFormer-S on the SWEC iEEG dataset.

from braindecode.models import MVPFormer

model = MVPFormer.from_pretrained("braindecode/mvpformer-s-pretrained")

The classification head is randomly initialised — fine-tune it, or call model.reset_head(n_outputs=...). Only the backbone is pre-trained.

Training data

The data that shaped these weights. The ids are the datasets: field of this card, which braindecode reads into model.card_data.datasets when you call from_pretrained (MOABB class names for MOABB datasets, OpenNeuro / NEMAR accessions, Hub ids otherwise). nm/on ids are datasets on NEMAR (on = NEMAR's copy of the OpenNeuro dataset with the same number). Rows marked private: are data the authors did not release; they are not in datasets:.

  • generative (next-segment contrastive) pretraining on 18 of the 68 SWEC subjects; 50 subjects held out
  • Sources: arXiv:2506.20354 (ICLR 2026) Section 5.1; arXiv:2506.20354 Section 4; https://github.com/IBM/multi-variate-parallel-transformer README, Checkpoints; arXiv:2506.20354 reproducibility statement (PDF text extraction).

Licence and intended use

Weights © IBM Corp. 2024-2025, released under Apache-2.0 (SPDX-License-Identifier: Apache-2.0) by https://github.com/IBM/multi-variate-parallel-transformer. Redistributed here unmodified in substance; only the parameter names were remapped onto braindecode's MVPFormer.

The authors attach the following restriction, reproduced verbatim from the upstream README:

This software may only be used for research. For other applications any liability is denied. In particular, the software must not be used for diagnostic purposes.

The paper's broader-impact statement adds: "both our model and our dataset are developed purely for research, and may not be used for any other purpose, especially diagnostic."

Citation

@article{carzaniga2025foundation,
  title={A foundation model with multi-variate parallel attention to generate neuronal activity},
  author={Carzaniga, Francesco and Hersche, Michael and Sebastian, Abu and Schindler, Kaspar and Rahimi, Abbas},
  journal={arXiv preprint arXiv:2506.20354},
  year={2025}
}

Cite Braindecode

If you use this braindecode port in scientific work, cite the software on Zenodo, doi:10.5281/zenodo.17699192, and the original model and datasets cited above.

@software{braindecode,
  author = {Aristimunha, Bruno and
            Guetschel, Pierre and
            Wimpff, Martin and
            Gemein, Lukas and
            Rommel, Cedric and
            Banville, Hubert and
            Sliwowski, Maciej and
            Wilson, Daniel and
            Brandt, Simon and
            Gnassounou, Théo and
            Paillard, Joseph and
            {Junqueira Lopes}, Bruna and
            Sedlar, Sara and
            Moreau, Thomas and
            Chevallier, Sylvain and
            Gramfort, Alexandre and
            Schirrmeister, Robin Tibor},
  title = {Braindecode: toolbox for decoding raw electrophysiological brain data
           with deep learning models},
  url = {https://github.com/braindecode/braindecode},
  doi = {10.5281/zenodo.17699192},
  publisher = {Zenodo},
  license = {BSD-3-Clause},
}
Downloads last month
5,682
Safetensors
Model size
73.6M params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for braindecode/mvpformer-s-pretrained

Finetunes
3 models

Paper for braindecode/mvpformer-s-pretrained