Dataset Card
LVIS-20k
A public computer-vision dataset of 19,808 images with 244,700 annotations across 1,203 classes, annotated as polygon and bounding box.
Pictograph Research · CC-BY-4.0 · Public · Updated 2026-08-08
Overview
This dataset holds 19,808 images annotated as polygon and bounding box across 1,203 classes. The most common auto-detected tags are playing, eating, person, eye-level view, top-down view, zoo. It is ready to fork, export in a dozen formats, or train a model from.
At a glance
| Metric | Value |
|---|---|
| Images | 19,808 |
| Annotations | 244,700 |
| Classes | 1,203 |
| Annotation types | polygon, bounding box |
| Folders | 1 |
| Storage | Standard |
| Visibility | Public |
What's in these images
Tags auto-detected across the dataset, most common first.
| Tag | Category | Images |
|---|---|---|
| playing | Attributes | 2,720 |
| eating | Attributes | 2,222 |
| person | Objects | 1,932 |
| eye-level view | Attributes | 1,486 |
| top-down view | Attributes | 1,381 |
| zoo | Scenes | 926 |
| driving | Attributes | 869 |
| child | Objects | 859 |
| walking | Attributes | 858 |
| bathroom | Scenes | 790 |
| cooking | Attributes | 779 |
| working | Attributes | 771 |
| sign | Objects | 766 |
| jumping | Attributes | 746 |
| ski | Objects | 729 |
Use this dataset
from pictograph import Client, Dataset
client = Client() # reads PICTOGRAPH_API_KEY
dataset: Dataset = client.datasets.get("LVIS-20k")
print(dataset.image_count, dataset.annotation_count)- Fork it into your own workspace with one click from its page.
- Pull it with the Pictograph SDK (shown above) or the CLI.
- Export to COCO, YOLO, Darwin V7, Pascal VOC, and more.
- Train a model directly from an export.
License
Released under CC-BY-4.0.
Creative Commons Attribution 4.0 Data / content
Allows any use, including commercial, as long as appropriate credit is given. A common choice for datasets.
Read the full Creative Commons Attribution 4.0 →More from Pictograph Research
Start building today
$5/mo free compute. No payment method required.