# jcjohnson/densecap

Dense image captioning in Torch

Repository: https://github.com/jcjohnson/densecap
Canonical: https://ross.abutalabs.com/products/densecap
Language: Jupyter Notebook
License: MIT
License Family: permissive
Last push: 2018-07-31T03:41:10+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": 3929, "days_push": 2955, "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 1596, forks 425 (observed 2026-08-28T04:05:09.145678+00:00)

## What it is
DenseCap is a Torch implementation of a fully convolutional localization network that performs dense captioning: detecting objects in images and generating natural-language descriptions for each region. It includes a pretrained model trained on Visual Genome, inference code for CPU/GPU, webcam demo, evaluation, and training code.

## Use cases
- generate captions for regions of an image
- detect and describe objects in photos
- run dense captioning on new images
- train a dense captioning model on Visual Genome
- evaluate dense captioning models
- run a live webcam captioning demo

## When to choose
- you need a research-grade dense captioning model with pretrained weights
- you want to reproduce or build on the CVPR 2016 DenseCap paper
- you can work with the legacy Torch (Lua) ecosystem

## When to avoid
- you need a maintained tool or modern framework support (Torch is deprecated)
- you want PyTorch or TensorFlow implementations
- you need production deployment with active support

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, deep-learning, computer-vision, image-processing, nlp
- domain: computer-vision, deep-learning, machine-learning, artificial-intelligence
- platform: python
- tags: dense-captioning, torch, object-detection, image-captioning, research-code, cvpr-2016, visual-genome, linux, macos, gpu

## Member repositories
- jcjohnson/densecap (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:09.145678+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 2026-08-30T03:54:03.242779+00:00, confidence not recorded.
  - readme: https://github.com/jcjohnson/densecap (fetched 2026-08-28T04:05:09.145678+00:00, sha 2e65afe670a7)
- Data as of 2026-08-30T08:39:29.467469+00:00.