skills.rest
nebius
nebiusOFFICIAL
Explore all skills by this author →
WebsiteGitHub

fiftyone

Deploy and manage FiftyOne for dataset curation and visualization in physical-AI workflows.

17|8|Updated Apr 7, 2026
Search Tags:#curation#visualization#robotics#dataset#fiftyone#physical-ai
One-click install
npx skills add https://github.com/nebius/nebius-physical-ai --skill fiftyone-nebius
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: fiftyone
Source: https://github.com/nebius/nebius-physical-ai/tree/main/skills/tools/fiftyone
Command: npx skills add https://github.com/nebius/nebius-physical-ai --skill fiftyone-nebius

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill addresses the complexity of managing, visualizing, and curating large-scale datasets for physical-AI workloads, removing the friction of manual data inspection.

Core Features & Use Cases

  • Dataset Curation: Perform real curation using FiftyOne Brain to compute uniqueness, similarity, and visual embeddings.
  • Visualization: Launch a dedicated interface to inspect augmented scenario variants and cluster near-duplicates.
  • Use Case: Robotics teams can use this to automatically filter out redundant synthetic data variants from a simulation run, ensuring only high-value, unique data is used for policy training.

Quick Start

Use the fiftyone skill to deploy the workbench tool and load your dataset for immediate visualization and curation.

Frequently Asked Questions about fiftyone

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

FAQPage Schema
How do I visualize and curate large-scale physical-AI datasets?▼

To visualize and curate physical-AI datasets, you can deploy a dedicated interface to inspect augmented scenario variants, compute visual embeddings, and cluster near-duplicates for high-value data extraction.

What is the best way to filter redundant synthetic data variants from simulation runs?▼

Filtering redundant synthetic data variants involves using embedding-based uniqueness and similarity computation to automatically identify and remove near-duplicates, ensuring only unique data is used for policy training.

Does this dataset visualization tool support deployment on Kubernetes infrastructure?▼

Yes, the dataset visualization tool integrates with both Kubernetes and local infrastructure, providing a unified interface for data factory reporting and model training preparation across your existing environments.

Can I use FiftyOne Brain to compute similarity for robotics datasets?▼

Yes, you can use FiftyOne Brain to perform real curation on robotics datasets by computing uniqueness, similarity, and visual embeddings to analyze and filter your physical-AI data.

Why should I use an embedding-based approach for dataset curation?▼

An embedding-based approach for dataset curation allows you to automatically identify high-value, unique data and cluster near-duplicates, removing the friction of manual data inspection for physical-AI workloads.

Related Skills

Explore More Data & Analytics Skills→
nebiusnebius
17

sim-to-real

Orchestrate robotics simulation, policy training, and evaluation workflows via declarative YAML.

Official
Advanced
zhuangbiaoweizhuangbiaowei
2

simpo-training

Optimize LLM alignment with reference-free SimPO preference optimization.

Community
Advanced
juanlamadrid20juanlamadrid20

databricks-unstructured-pdf-generation

Generate synthetic PDF documents with JSON metadata for RAG evaluation.

Community
Intermediate
niznik-devniznik-dev
16

create-inspect-task

Create custom inspect-ai evaluation tasks with experiment-guided and standalone modes.

Community
Intermediate