# fff-rs/juice

The Hacker's Machine Learning Engine

Repository: https://github.com/fff-rs/juice
Canonical: https://ross.abutalabs.com/products/fff-rs-juice
Language: Rust
License Family: other
Topics: machine-learning, framework, rust, agnostic, extinsible, juice, coaster, cuda, opencl, hacktoberfest
Last push: 2024-07-22T18:07:55+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3518, "days_push": 772, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1132, forks 75 (observed 2026-08-28T04:03:42.749507+00:00)

## What it is
Juice is a Rust machine learning framework ('The Hacker's Machine Learning Engine') built on the Coaster hardware abstraction layer, supporting CUDA, OpenCL, and native BLAS backends. The workspace includes the core framework, math abstractions, data preprocessing (greenglas), and a CLI for running examples like MNIST.

## Use cases
- train neural networks in rust
- run machine learning on gpu with cuda or opencl
- hardware-agnostic deep learning framework
- learn ml framework internals in rust
- run mnist training example from cli

## When to choose
- you want a Rust-native ML framework with pluggable GPU backends
- you need CUDA and OpenCL support from the same codebase
- you want to hack on or extend an ML engine

## When to avoid
- you need production-grade training with broad model support - PyTorch or TensorFlow are far more mature
- you cannot set up CUDA/cuDNN, which is required for examples
- you need active community support - development is slow

## Facets
- artifact type: framework
- maturity: maintenance
- function: machine-learning, deep-learning, gpu-computing, cli
- domain: machine-learning, deep-learning, developer-tools
- platform: rust, cli
- tags: cuda, opencl, blas, hardware-agnostic, capnproto, neural-networks, linux, gpu

## Member repositories
- fff-rs/juice (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:42.749507+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 2026-08-30T06:37:57.132246+00:00, confidence not recorded.
  - readme: https://github.com/fff-rs/juice (fetched 2026-08-28T04:03:42.749507+00:00, sha c5a56f6dd41a)
- Data as of 2026-08-30T08:39:29.467469+00:00.