# xmartlabs/Bender

Easily craft fast Neural Networks on iOS! Use TensorFlow models. Metal under the hood.

Repository: https://github.com/xmartlabs/Bender
Canonical: https://ross.abutalabs.com/products/bender
Homepage: https://xmartlabs.github.io/Bender/
Language: Swift
License: MIT
License Family: permissive
Topics: machine-learning, neural-networks, metal, apple, iphone, ios, convolutional-neural-networks, deep-learning, swift, deep-neural-networks, residual-networks
Last push: 2023-11-07T10:28:36+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": 3409, "days_push": 1030, "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 1802, forks 89 (observed 2026-08-28T04:05:38.136334+00:00)

## What it is
Bender is a Swift library that provides an abstraction layer over Apple's MetalPerformanceShaders for defining and running neural networks on iOS. It supports importing pretrained TensorFlow models and executes inference directly on the GPU.

## Use cases
- run pretrained neural networks on iphone
- run tensorflow models on ios
- on-device image style transfer
- pose detection in an ios app
- run inference on gpu with metal
- add custom layers coreml doesn't support

## When to choose
- you need custom or unsupported layers that CoreML can't run
- you want full GPU control via MetalPerformanceShaders instead of CoreML's throttling
- you have frozen TensorFlow graphs to run on iOS

## When to avoid
- you need a broadly supported, actively updated framework - CoreML or TensorFlow Lite are better maintained
- your target is Android or cross-platform
- you need training on device - Bender only runs inference

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, llm-inference
- domain: machine-learning, mobile-development, computer-vision
- platform: -
- tags: metal, metalperformanceshaders, tensorflow-import, on-device-inference, neural-networks, gpu-acceleration, ios, swift, mobile, gpu

## Member repositories
- xmartlabs/Bender (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:38.136334+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 2026-08-30T03:22:03.385041+00:00, confidence not recorded.
  - readme: https://github.com/xmartlabs/Bender (fetched 2026-08-28T04:05:38.136334+00:00, sha c208bc561bc7)
  - homepage: https://xmartlabs.github.io/Bender/ (fetched 2026-08-29T11:00:50.352453+00:00, sha 0f7b0a6121e8)
- Data as of 2026-08-30T08:39:29.467469+00:00.