|
--- |
|
license: mit |
|
task_categories: |
|
- reinforcement-learning |
|
--- |
|
|
|
# Dark Key-Door |
|
This repository contains the Dark Key-Door datasets as used in ["Retrieval-augmented Decision Transformer: External Memory for In-context RL"](https://huggingface.co/papers/2410.07071): |
|
- Datasets for 3 grid-sizes: 10x10, 20x20, 40x20. |
|
- Every split contains 100 `.npz` files, corresponing to 100 start-goal position combinations. |
|
- Each file contains 100K transitions. |
|
- Filenames have the format `START_KEY_DOOR.npz`. E.g., for `[0,0]_[0,5]_[0,7].npz`, `[0.0]` is the start position, the key is located at `[0,5]`, and the door is found at `[0,7]`. |
|
|
|
Download the dataset using: |
|
``` |
|
huggingface-cli download ml-jku/dark_keydoor --local-dir=./dark_keydoor --repo-type dataset |
|
``` |
|
For dataloading we refer to our Github repository: https://github.com/ml-jku/RA-DT |
|
|
|
Citation: |
|
``` |
|
@article{schmied2024retrieval, |
|
title={Retrieval-Augmented Decision Transformer: External Memory for In-context RL}, |
|
author={Schmied, Thomas and Paischer, Fabian and Patil, Vihang and Hofmarcher, Markus and Pascanu, Razvan and Hochreiter, Sepp}, |
|
journal={ArXiv}, |
|
year={2024} |
|
} |
|
``` |
