This site hosts benchmarked results of Spiking Neural Networks (SNNs) from leading research papers. The goal is to track the leading approaches and the current state of the art in spiking neural networks.
The site is live at snnbench.com (also snnbenchmarks.com). If you would like to contribute, please reach out — contact details are on my personal site daniellong.co.
index.html— the benchmarks table, populated byscript.jsfromdata/.background.html+background/— explanatory pages about SNNs.contact.html— contact form (web3forms).data/papers.json,data/datasets.json— the benchmark data.css/— stylesheets.partials/— shared<head>and<header>markup (see Build).
The shared <head> and <header> are defined once in partials/ and injected into
every page by build.js. The HTML files are committed fully-rendered, so the site works
when served directly with no build step.
After editing anything in partials/, or the per-page titles/descriptions in the PAGES
config in build.js, re-sync all pages with:
node build.jsThis rewrites the <head> and <header> of every page listed in PAGES. It is
idempotent — running it without changes leaves the files untouched.
python3 -m http.server 8000
# then open http://localhost:8000/index.html