# vdumoulin/conv_arithmetic

A technical report on convolution arithmetic in the context of deep learning

Repository: https://github.com/vdumoulin/conv_arithmetic
Canonical: https://ross.abutalabs.com/products/conv_arithmetic
Language: TeX
License: MIT
License Family: permissive
Archived: true
Last push: 2023-06-08T21:06:47+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 3843, "days_push": 1182, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 14690, forks 2307 (observed 2026-08-28T04:11:07.457931+00:00)

## What it is
A technical report and tutorial repository explaining convolution arithmetic for deep learning, with animated visualizations of convolutions, transposed convolutions, and dilated convolutions. It accompanies the arXiv paper 'A guide to convolution arithmetic for deep learning' and includes code to regenerate the animations.

## Use cases
- understand how padding and strides affect convolution output shapes
- learn transposed convolution for upsampling in CNNs
- visualize dilated convolutions
- find a reference for convolution output size formulas
- teaching material for deep learning courses

## When to choose
- you need clear visual explanations of convolution operations
- you want a citable technical reference on convolution arithmetic
- you are learning or teaching CNN fundamentals

## When to avoid
- you need a production convolution library or implementation
- you want up-to-date coverage of modern architectures beyond convolutions

## Facets
- artifact type: learning-resource
- maturity: stable
- function: machine-learning, deep-learning, data-visualization, documentation
- domain: deep-learning, machine-learning, tutorials, education
- platform: python, cross-platform
- tags: convolutional-neural-networks, technical-report, animations, transposed-convolution, dilated-convolution, latex

## Member repositories
- vdumoulin/conv_arithmetic (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:07.457931+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 2026-08-29T17:07:19.290255+00:00, confidence not recorded.
  - readme: https://github.com/vdumoulin/conv_arithmetic (fetched 2026-08-28T04:11:07.457931+00:00, sha 0fb2546e040e)
- Data as of 2026-08-30T08:39:29.467469+00:00.