Organization
Location
CH
Badges
Activity
Challenge Categories
Challenges Entered
Completed
MineRL Labs
Learning From Human-Feedback
Latest submissions
AIcrowd
IGLU Team
Interactive embodied agents for Human-AI collaboration
Latest submissions
See All| failed | 196477 | ||
| failed | 196475 |
ASCII-rendered single-player dungeon crawl game
Latest submissions
Parametrix.ai
MIT
THU_SIGS
Latest submissions
MineRL Labs - Carnegie Mellon University
Training sample-efficient agents in Minecraft
Latest submissions
OpenAI
Measure sample efficiency and generalization in reinforcement learning using procedurally generated environments
Latest submissions
SBB
Deutsche Bahn
SNCF
Multi-Agent Reinforcement Learning on Trains
Latest submissions
Use an RL agent to build a structure with natural language inputs
Latest submissions
See All| failed | 196477 | ||
| failed | 196475 | ||
| graded | 194819 |
Language assisted Human - AI Collaboration
Latest submissions
| Participant | Rating |
|---|
| Participant | Rating |
|---|
NeurIPS 2022 IGLU Challenge
Baselines code
Over 3 years agoHello,
Yes, there is a trained baseline to be released soon. We expect to release it by the end of the warm-up round.
In the meantime, I suggest you get familiar with the idea behind our baseline by studying the RL env paper: https://arxiv.org/pdf/2206.00142.pdf
And the WIP baseline codebase: https://github.com/iglu-contest/iglu-2022-rl-baseline
Please note that since this code is not been officially released yet, it is not guaranteed to work properly.
Definition of action space
Over 3 years agoHello,
Sorry for the long reply. The
placementaction key controls placing/breaking blocks in the gridworld. The mapping is the following: 0 - noop, 1 - place block, 2 - break block.I hope this helps.