OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Neurorvq

Neurorvq is a Python research implementation of NeuroRVQ, a multi-scale tokenizer and family of biosignal foundation models. Its tokenizers convert EEG, EMG, ECG and PPG signals into discrete representations; associated models use masked-token prediction to learn biosignal representations.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Neurorvq — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use NeuroRVQ when a research workflow needs biosignal tokenization, reconstruction examples, or pretrained models for representation-learning experiments.

Who it is for

  • Researchers studying biosignal tokenization and generative foundation models
  • Developers exploring EEG, ECG, EMG or PPG representation learning
  • Brain-computer-interface researchers investigating downstream use of learned biosignal representations

Documented uses

  • Convert multivariate biosignal time series into discrete, multi-scale tokens.
  • Inspect reconstruction examples for supported biosignal modalities.
  • Load pretrained tokenizer and foundation-model weights for research.
  • Explore fine-tuning foundation models, following the repository’s data instructions for custom datasets.
Before you choose

Requirements & limits

  • The project supplies code and pretrained model files rather than a documented hosted service.
  • The documented setup uses Conda, Python 3.10 and the repository requirements file.
  • The README lists tokenizer variants for EEG, EMG, ECG and PPG, but foundation-model variants for EEG, EMG and ECG.
  • Custom-dataset fine-tuning depends on following the repository’s data instructions.
  • No license is listed in the repository metadata.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install NeuroRVQ and download EEG model files

Prepare a Python 3.10 Conda environment with the repository requirements and download the EEG tokenizer, foundation model, and sample signal.

Before you start

  • Conda
  • The NeuroRVQ repository cloned or otherwise available locally

Steps

  1. Create and activate the environment: conda create -n neurorvq python=3.10 and conda activate neurorvq.

  2. From the repository directory, install dependencies with the command below.

    Documented command
    pip install -r requirements.txt

  3. In Python, use hf_hub_download from huggingface_hub to download pretrained_models/tokenizers/NeuroRVQ_EEG_tokenizer_v1.pt, pretrained_models/foundation_models/NeuroRVQ_EEG_foundation_model_v1.pt, and example_files/eeg_sample/example_eeg_file.xdf from ntinosbarmpas/NeuroRVQ, setting local_dir="./" as in th

Expected outcome

The environment has the repository requirements installed, with the EEG tokenizer, foundation model, and example EEG file downloaded to the selected local directory.

Guide limitations

  • The repository documentation gives no detailed dependency or hardware specifications beyond Python 3.10 and installing requirements.txt.
  • The download example uses local_dir="./"; model-loading examples expect model files under ./pretrained_models/..., so ensure the downloaded files are placed at the paths used by those examples before loading.
  • Custom-dataset preparation is not covered by this setup recipe.
Documentation 1 ↗Documentation 2 ↗
Source-based practical guide

Run the EEG tokenizer reconstruction example

Load the EEG tokenizer and request its documented reconstruction example and plots.

Before you start

  • NeuroRVQ environment with requirements.txt installed
  • EEG tokenizer file available at ./pretrained_models/tokenizers/NeuroRVQ_EEG_tokenizer_v1.pt

Steps

  1. In the repository environment, import load_neurorqv_tokenizer from inference.run.NeuroRVQ_EEG_tokenizer_example.

  2. Call load_neurorqv_tokenizer(run_example=True, plot_results=True, verbose=True, model_path='./pretrained_models/tokenizers/NeuroRVQ_EEG_tokenizer_v1.pt').

  3. Use the example function’s documented options to inspect reconstruction results.

Expected outcome

The example is configured to run the EEG tokenizer reconstruction and display plots.

Guide limitations

  • The documented call uses the example function; the README directs users to inspect that function to load and use the tokenizer in other workflows.
  • The repository text does not specify expected numerical results or verify execution in a particular environment.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
KonstantinosBarmpas/NeuroRVQ · Official homepage title · retained in its source language
How we review sourcesSuggest a correction