# yosinski/deep-visualization-toolbox

DeepVis Toolbox

Repository: https://github.com/yosinski/deep-visualization-toolbox
Canonical: https://ross.abutalabs.com/products/deep-visualization-toolbox
Homepage: http://yosinski.com/deepvis
Language: Python
License: MIT
License Family: permissive
Last push: 2020-01-13T19:32:50+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": 4148, "days_push": 2424, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4051, forks 928 (observed 2026-08-28T04:08:33.822644+00:00)

## What it is
A GUI toolbox for visualizing and understanding deep neural networks, showing per-unit activations, backprop/deconv, and regularized-optimization visualizations of what individual neurons respond to. Bundled with CaffeNet, GoogLeNet, and SqueezeNet models and described in an ICML 2015 Deep Learning Workshop paper.

## Use cases
- visualize what individual neurons in a CNN have learned
- run images through a network and inspect activations layer by layer
- generate synthetic images that maximally activate a neuron
- understand why a neural network classifies an image a certain way
- teach deep learning interpretability concepts with a live demo
- visualize network activations from a webcam feed

## When to choose
- you use Caffe-based CNN models and want interactive neuron-level visualization
- you are studying or teaching neural network interpretability
- you want to reproduce the Yosinski et al. 2015 deep visualization experiments

## When to avoid
- you work with PyTorch or TensorFlow models, since the toolbox is built around Caffe
- you need actively maintained tooling - the project has not seen releases since 2020
- you need modern interpretability methods like Grad-CAM, SHAP, or attention visualizations

## Facets
- artifact type: application
- maturity: maintenance
- function: machine-learning, data-visualization, deep-learning, gui
- domain: deep-learning, machine-learning, computer-vision, data-visualization
- platform: python
- tags: neural-network-visualization, caffe, interpretability, deconvnet, research-tool, linux, macos, desktop

## Member repositories
- yosinski/deep-visualization-toolbox (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:33.822644+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 2026-08-29T18:23:40.605549+00:00, confidence not recorded.
  - readme: https://github.com/yosinski/deep-visualization-toolbox (fetched 2026-08-28T04:08:33.822644+00:00, sha 43db654c8ba9)
  - homepage: http://yosinski.com/deepvis (fetched 2026-08-29T09:16:01.796860+00:00, sha 3958cf4d30cb)
- Data as of 2026-08-30T08:39:29.467469+00:00.