Neuromorphic Toolchain
Spiking Neural Network Toolchain

Neuromorphic AI Development Tools

Neuromorphic computing trades dense matrix multiplication for sparse, event-driven spikes — networks of artificial neurons that fire only when they need to, on hardware built to exploit that sparsity. Five tools currently anchor the toolchain, from converting an existing PyTorch model down to a spiking form, to simulating custom neuron models, to shipping onto neuromorphic silicon. This is a field guide to what each one is for.

01 — The connective layer

NIR sits at the center of the toolchain

Neuromorphic Intermediate Representation (NIR) is the hardware-agnostic graph format that lets a model move between tools — export once from any NIR-producing tool, deploy to any NIR-compatible simulator or hardware target, the way ONNX decouples a model from any single framework.

NIR interchange format NeuroCUDA → GPU · CPU · Loihi 2 sim MetaTF → BrainChip Akida chip snnTorch trains SNNs from scratch GeNN → generated CUDA / GPU
NeuroCUDA exports bit-exact NIR for graphs like ResNet-18 — the other tools read and write the same format.
02 — A typical workflow

Where each tool sits in the pipeline

Most projects touch two or three of these tools, not all five. The path usually runs from a source model or a from-scratch design, through NIR, and out to a simulator or physical neuromorphic chip.

STEP 1

Get a spiking model

Convert a trained PyTorch model, or design a spiking architecture from first principles.

NeuroCUDA · snnTorch
STEP 2

Exchange via NIR

Serialize the network graph to a hardware-agnostic, NIR format any downstream tool can read.

NIR
STEP 3

Simulate

Run the network at scale, or model custom neurons for neuroscience research, on GPU-generated CUDA code.

GeNN
STEP 4

Deploy to silicon

Ship to neuromorphic hardware with on-chip learning for a commercial product.

MetaTF → Akida
03 — The tools

Five tools, five jobs

NeuroCUDA

Compiler / Converter

A PyTorch-to-SNN compiler with GPU, CPU, and Loihi 2 simulator backends — takes a model you've already trained and converts it into a spiking form, ready to target several backends from one artifact.

MIT license
$ pip install neurocuda
Best forConverting existing trained PyTorch models.

snnTorch

Training library

A PyTorch library for training spiking neural networks from scratch, using surrogate-gradient backpropagation through time to get around the non-differentiability of a spike.

Best forResearchers designing SNN architectures from first principles.

GeNN

Simulator

A GPU-accelerated SNN simulator that generates optimized CUDA code directly from high-level neuron descriptions, rather than interpreting a generic graph at runtime.

Best forComputational-neuroscience simulation of custom neuron models.

NIR — Neuromorphic Intermediate Representation

Interchange format

The hardware-agnostic graph format connecting the other four tools — the neuromorphic equivalent of ONNX. Export a model once, and deploy it to any NIR-compatible simulator or hardware target. NeuroCUDA, for instance, exports bit-exact NIR for graphs such as ResNet-18.

Best forMoving a model between training, simulation, and hardware without re-authoring it per target.

MetaTF — BrainChip

Commercial toolchain

A TensorFlow model converter for deployment on BrainChip's Akida neuromorphic processor, with support for on-chip learning rather than a purely fixed, pre-trained model.

Best forCommercial product development on Akida-based hardware.