Wildworld Dataset

This loads and works with the WildWorld dataset, a massive 108M+ frame collection from a photorealistic ARPG that's designed for training world models that predict what happens next in a game. Every frame comes with skeleton data, character states like HP and stamina, depth maps, camera poses, and labels for 450+ actions. The skill provides loader utilities for iterating through clips and a PyTorch Dataset class for feeding sequences into models. Worth noting the actual dataset hasn't been released yet as of the paper publication, so this is more of a reference implementation waiting on the data drop. If you're working on video prediction or action-conditioned generation for games, the annotation density here is exceptional once it ships.

skills.sh installs: 739
GitHub stars: 61
last updated: Jul 14, 2026

View on GitHub

Install

$ npx -y skills add aradotso/trending-skills --skill wildworld-dataset --agent claude-code

Installs into .claude/skills of the current project.

Files

WildWorld Dataset Skill

Skill by ara.so — Daily 2026 Skills collection.

What WildWorld Is

WildWorld is a large-scale action-conditioned world modeling dataset automatically collected from a photorealistic AAA action role-playing game (ARPG). It is designed for training and evaluating dynamic world models — generative models that predict future game states given past observations and player actions.

Key Statistics

Property	Value
Total frames	108M+
Actions	450+ semantically meaningful
Monster species	29
Player characters	4
Weapon types	4
Distinct stages	5
Max clip length	30+ minutes continuous

Per-Frame Annotations

Every frame includes:

Character skeletons — joint positions for player and monsters
Actions & states — HP, animation state, stamina, etc.
Camera poses — position, rotation, field of view
Depth maps — monocular depth for each frame
Hierarchical captions — action-level and sample-level natural language descriptions