taisui 0.1.1

Third-generation Artificial Intelligence SNN Universal Implementation

TAISUI

English | 中文(Chinese)

TAISUI is a low-power, large-scale spiking neural network training and inference framework. It preserves minimal energy cosumption while providing the superior learning abilities of large spiking neural network.

The vision of TAISUI is to decode human intelligence and the mechanisms of its evolution, and to provide support for the development of brain-inspired intelligent agents in a future society where humans coexist with artificial intelligence.

• Documentation
• Source

Install

TAISUI offers two installation methods. Running the following command in your terminal will install the project:

Install from PyPI:

pip install taisui

Install from GitHub:

1. Download or clone TAISUI from github

git clone https://github.com/taisui/taisui.git

2. Enter the folder of TAISUI and install braincog locally with setuptools

cd taisui
python setup.py install

Quickstart

The code style of TAISUI is consistent with Pytorch, allowing you to build spiking neural networks with simple code:

from taisui.base.neuron import LIFNode
import torch

x = torch.randn(2, 3, 5, 5)

net = torch.nn.Sequential(
    nn.Conv2d(1, 32, kernel_size=3),
    LIFNode(),
    nn.Flatten(),
    nn.Linear(54, 1)
)

y = net(x)

Development plans

TAISUI is still under active development:

• Large-scale deep spiking neural network training and inference
• Ultra-low energy consumption sparse spiking neural network computing
• Brain-inspired learning algorithm support
• Bionic neural network sparse structure support

Cite

If you are using TAISUI, please consider citing it as follows:

@misc{TAISUI,
    title = {TAISUI},
    author = {Lei, Yunlin and Gao, Lanyu and Yang, Xu and other contributors},
    year = {2024},
    publisher = {GitHub},
    journal = {GitHub repository},
    howpublished = {\url{https://github.com/taisui/taisui}},
}

About

Beijing Institute of Technology AETAS Laboratoryare the main developers of TAISUI.