{
  "id": 9012588,
  "title": "SpikeForge grew into a toolkit, a model hub and a dashboard",
  "url": "https://urgent.news/2026/09/21/spikeforge-grew-into-a-toolkit-a-model-hub-and-a-dashboard",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-21T21:21:52.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/w4ffl35/spikeforge-grew-into-a-toolkit-a-model-hub-and-a-dashboard-37k3"
  },
  "original_language": "en",
  "account": "SpikeForge, a Python toolkit for developing and testing spiking neural networks, has evolved into a comprehensive toolkit, model hub, and dashboard with the recent release cycle. This expansion involved the separation of the core toolkit, target adapters, model hub, and SpikeForge Dashboard into distinct packages and applications.\n\nThe core project has released versions 0.4.0 and 0.5.0, followed by several smaller package releases to enhance the ecosystem's usability. Notably, the event path now includes a real train/test split, while quantization work has been integrated into drift checks. Shared logging has been sourced from capsize-commons, and the CI path has been optimized with a fast leaf lane for non-workspace components.\n\nThe package layout now offers a clear map of the work, with spikeforge focusing on training and experiment workflow, spikeforge-targets handling target-specific conversions, spikeforge-hub managing the offline-first model catalogue, and spikeforge-dashboard providing the web dashboard and desktop application. This separation offers practical benefits, allowing users to install only the necessary components based on their needs. For example, those experimenting in Python need not install the desktop app, while those browsing models do not require the server.\n\nThe stack consists of Python, PyTorch, and snnTorch around the core, with a web dashboard and desktop application built on top. The model hub is designed to be cache-friendly, enabling users to inspect and utilize the catalogue without converting every experiment into a network operation. The SpikeForge Dashboard provides a visual entry point and allows users to inspect experiments beyond their final numbers.\n\nFurther information on package links, release notes, and desktop releases can be found on the project page (spikeforge.net) and on itch.io. The source code is available on GitHub.\n\nFor a hands-on experience, users are encouraged to install the core package, select an example, and run a small experiment using a single dataset and a brief training window. This experiment should then progress through the target adapter and be saved in the hub catalogue, focusing on understanding the path from an experiment to a repeatable artefact. A follow-up article will delve into a small event-driven classifier, and the four-week overview will cover additional details from the same period.",
  "summary": "SpikeForge is my Python toolkit for building and testing spiking neural networks. It stopped behaving like one big project during this release cycle: the core toolkit, target adapters, model hub and SpikeForge Dashboard all moved forward as separate packages and applications. The core project released 0.4.0 and 0.5.0, then kept publishing the smaller package releases that make the ecosystem…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}