mrwabbit's picture
Upload README.md with huggingface_hub
28f3e5d verified
|
raw
history blame contribute delete
2 kB
metadata
language: en
license: mit
library_name: pytorch
tags:
  - spiking-neural-network
  - neuromorphic
  - surrogate-gradient
  - benchmark
  - catalyst
  - nmnist
datasets:
  - n-mnist
metrics:
  - accuracy
model-index:
  - name: Catalyst NMNIST SNN Benchmark (N3)
    results:
      - task:
          type: image-classification
          name: Neuromorphic Digit Classification
        dataset:
          name: Neuromorphic MNIST (N-MNIST)
          type: n-mnist
        metrics:
          - name: Float Accuracy (N3)
            type: accuracy
            value: 99.2

Catalyst NMNIST SNN Benchmark (N3)

Convolutional Spiking Neural Network for neuromorphic digit classification on N-MNIST.

Model Description

  • Architecture (N3): Conv2d → LIF → 10
  • Neuron model: Leaky Integrate-and-Fire (LIF) with convolutional feature extraction
  • Training: Surrogate gradient BPTT, fast-sigmoid surrogate (scale=25)
  • Hardware target: Catalyst N3 neuromorphic processor

Results

Metric Value
Float accuracy 99.2%
Parameters 691,210

Reproduce

git clone https://github.com/catalyst-neuromorphic/catalyst-benchmarks.git
cd catalyst-benchmarks
pip install -e .
python nmnist/train.py --device cuda:0

Links

Citation

@misc{catalyst-benchmarks-2026,
  author = {Shulayev Barnes, Henry},
  title = {Catalyst Neuromorphic Benchmarks},
  year = {2026},
  url = {https://github.com/catalyst-neuromorphic/catalyst-benchmarks}
}