fiftyone-dataset-import

Automate dataset imports into FiftyOne with automatic format detection.

37|8|Updated Dec 18, 2025
One-click install
npx skills add https://github.com/voxel51/fiftyone-skills --skill fiftyone-dataset-import
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: fiftyone-dataset-import
Source: https://github.com/voxel51/fiftyone-skills/tree/main/skills/fiftyone-dataset-import
Command: npx skills add https://github.com/voxel51/fiftyone-skills --skill fiftyone-dataset-import

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates importing diverse datasets into FiftyOne with automatic format detection, removing the manual guesswork and setup typically required to prepare data for FiftyOne.

Core Features & Use Cases

  • Auto-detects media types (images, videos, point clouds) and label formats (COCO, YOLO, VOC, KITTI) from local folders or Hugging Face Hub datasets.
  • Supports multimodal grouped datasets and scene-based structures for streamlined ingestion of complex data.
  • Provides guided import flows using FiftyOne's APIs, with prompts for user confirmation before creating datasets.
  • Includes practical examples for common workflows, such as importing a COCO dataset or loading a HF Hub dataset with preserved structure.

Quick Start

Scan a local directory or Hugging Face Hub repository and import the dataset into FiftyOne. Example: Import a dataset located at /path/to/dataset into FiftyOne.

Frequently Asked Questions about fiftyone-dataset-import

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I import a dataset into FiftyOne without knowing its format?

You can import a dataset into FiftyOne without specifying the format by using automatic detection. The Skill scans local directories or HuggingFace Hub datasets to identify media types and label formats like COCO or YOLO automatically.

Can I load HuggingFace Hub datasets into FiftyOne with their original structure?

Yes, you can load HuggingFace Hub datasets into FiftyOne with their original structure preserved. The Skill handles downloading and ingestion while supporting multimodal groupings and scene-based structures.

Does FiftyOne support importing multimodal datasets containing images, videos, and point clouds?

FiftyOne supports importing multimodal datasets containing images, videos, and point clouds. The Skill detects these diverse media types and loads them as grouped datasets to streamline ingestion of complex data.

Do I need to manually install dependencies to load different dataset formats in FiftyOne?

You do not need to manually install dependencies to load different dataset formats in FiftyOne. The Skill installs necessary format-specific dependencies on demand during the import process.

What is the best way to import a COCO dataset into FiftyOne?

The best way to import a COCO dataset into FiftyOne is to point the Skill at the local directory. It detects the COCO format, confirms the import parameters with you, and loads the data via FiftyOne's import operators.