# HuwCampbell/grenade

Deep Learning in Haskell

Repository: https://github.com/HuwCampbell/grenade
Canonical: https://ross.abutalabs.com/products/grenade
Language: Haskell
License: BSD-2-Clause
License Family: permissive
Topics: machine-learning, deep-neural-networks, haskell, deep-learning, generative-adversarial-networks, convolutional-neural-networks
Last push: 2023-12-08T21:23:07+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 3723, "days_push": 999, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1449, forks 81 (observed 2026-08-28T04:04:45.588915+00:00)

## What it is
Grenade is a composable, dependently typed deep learning library for Haskell that lets users specify neural network architectures at the type level, including convolutional, fully connected, and recurrent layers like LSTM. Networks are defined as heterogeneous lists of layers with shape-checked data flow, requiring minimal term-level code.

## Use cases
- train a CNN for MNIST digit classification in Haskell
- define an LSTM recurrent network for character-level text generation
- specify a neural network architecture with compile-time shape checking
- experiment with generative adversarial networks
- build deep learning models without leaving the Haskell ecosystem

## When to choose
- you are a Haskell developer wanting type-safe neural network definitions
- you want compile-time verification of layer shapes and data dimensions
- you need a concise declarative way to specify CNNs or RNNs

## When to avoid
- you need GPU acceleration or production-scale training performance
- you want a large ecosystem of pretrained models and tooling
- you are not working in Haskell

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning
- domain: machine-learning, deep-learning
- platform: cross-platform
- tags: neural-networks, cnn, rnn, lstm, gan, dependently-typed, haskell

## Member repositories
- HuwCampbell/grenade (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:45.588915+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:35:58.630953+00:00, confidence not recorded.
  - readme: https://github.com/HuwCampbell/grenade (fetched 2026-08-28T04:04:45.588915+00:00, sha d703d701c1ff)
- Data as of 2026-08-30T08:39:29.467469+00:00.