tylin/coco-caption

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1224 stars · 544 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

The official evaluation code for the Microsoft COCO image captioning benchmark, implementing metrics such as BLEU, METEOR, ROUGE-L, CIDEr, and SPICE. It includes annotation files, demo scripts, and a Python package (pycocoevalcap) for scoring generated captions against reference captions.

Use cases

• evaluate image captioning model outputs on COCO
• compute BLEU, METEOR, ROUGE-L, CIDEr, and SPICE scores for captions
• benchmark caption generation against MS COCO 2014 validation set
• score machine-generated captions against human references
• run caption evaluation in a research pipeline

When to choose

• you need standard COCO caption metrics for a paper or benchmark
• you are evaluating image captioning or multimodal generation models
• you want the official reference implementation of CIDEr and SPICE

When to avoid

• you need Python 3 support out of the box (code targets Python 2.7 and Java 1.8)
• you are evaluating tasks other than image captioning
• you want a maintained library with active development

Facets

library · maturity maintenance

nlp machine-learning benchmarking computer-vision machine-learning python image-captioning evaluation-metrics coco-dataset bleu cider spice meteor rouge natural-language-processing linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="tylin/coco-caption")