# hannes-brt/hebel

GPU-Accelerated Deep Learning Library in Python

Repository: https://github.com/hannes-brt/hebel
Canonical: https://ross.abutalabs.com/products/hebel
Language: Python
License: GPL-2.0
License Family: copyleft
Last push: 2020-12-29T05:32:20+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 4669, "days_push": 2073, "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 1169, forks 116 (observed 2026-08-28T04:03:51.080695+00:00)

## What it is
Hebel is a GPU-accelerated deep learning library in Python built on PyCUDA, implementing feed-forward neural networks for classification and regression with SGD, momentum, dropout, and regularization. The author no longer actively develops it and recommends Chainer instead.

## Use cases
- train feed-forward neural networks on GPU with CUDA
- classify MNIST digits with a shallow neural net
- run regression with deep learning in Python
- experiment with dropout and Nesterov momentum SGD
- train multi-task neural network models

## When to choose
- you need a lightweight CUDA-based neural network library with YAML-driven training configs
- you are maintaining legacy code already built on Hebel

## When to avoid
- you need an actively maintained deep learning framework
- you need convolutional networks, autoencoders, or modern architectures
- you lack an NVIDIA GPU with CUDA and PyCUDA

## Facets
- artifact type: library
- maturity: abandoned
- function: deep-learning, machine-learning, gpu-computing
- domain: deep-learning, machine-learning, gpu-computing
- platform: python, windows
- tags: cuda, pycuda, neural-networks, feedforward-networks, sgd, dropout, linux

## Member repositories
- hannes-brt/hebel (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:51.080695+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 2026-08-30T06:28:41.624759+00:00, confidence not recorded.
  - readme: https://github.com/hannes-brt/hebel (fetched 2026-08-28T04:03:51.080695+00:00, sha 0d050bf7abad)
  - registry_pypi: https://pypi.org/pypi/hebel/json (fetched 2026-08-29T12:34:41.441379+00:00, sha 7aca3d019adf)
- Data as of 2026-08-30T08:39:29.467469+00:00.