ferromorphic
v0.23.2 ExperimentalNeuromorphic computing in pure Rust, the open corpus in one library, every module checked against a closed form and mutation-audited: neurons from Hodgkin-Huxley to resonate-and-fire, STDP to three-factor and dendritic learning, surrogate gradients, BPTT, e-prop, exact spike-time gradients, EventProp, the spike response model, online training through time, diagonal state-space models, feedback alignment, DECOLLE, forward-forward, FORCE, ANN-to-SNN conversion, spiking convolutions, attention, reservoirs, the NEF, VSA, sparse distributed representations, sparse coding, spiking QUBO, oscillators, ring attractors, neural fields, predictive coding, a cerebellum, grid cells, delays and polychronous groups, spike-train metrics, spiking graphs, cochlea, olfaction, touch, proprioception, ECG/EMG, localisation, event vision, hardware models, NIR, event-camera readers, NeuroBench metrics, and a joules ledger. Zero dependencies, wasm-clean, deterministic by seed.
Quick Verdict
- โActively maintained (updated 8d ago)
- !Pre-1.0: API may have breaking changes
- โTiny footprint (3.6MB, 0 deps)
- โPermissive license (Apache-2.0)
Security
Deep Insights
428 downloads in the last 30 days (14/day avg).
The primary maintainer publishes 66 crates. This suggests deep Rust expertise and long-term commitment to the ecosystem.
ferromorphic has no external dependencies โ minimal supply chain risk and fast compile times.
3.6MB is relatively large. May impact download and compile times.
Health Breakdown
Recency, release consistency, active ratio
Yanked ratio, deps, size, maturity, features
Reverse deps, ownership, ecosystem
Downloads, momentum, growth trend
Docs, repo, license, metadata