SpikeForge Tutorial: Building an Event-Driven Classifier with Train/Test Split

SpikeForge is an open-source Python toolkit built on PyTorch and snnTorch for constructing event-driven neural network experiments. A recent release introduced a formal train/test split along the event path, making classifier experiments more reproducible and transparent. The recommended workflow involves loading a supported dataset, converting samples into event representations, selecting a compact network, and saving configuration details alongside results. Keeping experiments small and rerunnable helps isolate which settings influenced outcomes, a practice the project encourages through its modular package structure. SpikeForge allows users to swap models or add target adapters without rewriting data loaders or training loops, and finished experiments can be published to the SpikeForge model hub with versioning and provenance metadata.