Metadata-Version: 2.1
Name: xminigrid
Version: 0.4.0
Summary: JAX-accelerated meta-reinforcement learning environments inspired by XLand and MiniGrid
Author-email: Alexander Nikulin <a.p.nikulin@tinkoff.ai>
License:                                  Apache License
                                   Version 2.0, January 2004
                                http://www.apache.org/licenses/
        
           TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
        
           1. Definitions.
        
              "License" shall mean the terms and conditions for use, reproduction,
              and distribution as defined by Sections 1 through 9 of this document.
        
              "Licensor" shall mean the copyright owner or entity authorized by
              the copyright owner that is granting the License.
        
              "Legal Entity" shall mean the union of the acting entity and all
              other entities that control, are controlled by, or are under common
              control with that entity. For the purposes of this definition,
              "control" means (i) the power, direct or indirect, to cause the
              direction or management of such entity, whether by contract or
              otherwise, or (ii) ownership of fifty percent (50%) or more of the
              outstanding shares, or (iii) beneficial ownership of such entity.
        
              "You" (or "Your") shall mean an individual or Legal Entity
              exercising permissions granted by this License.
        
              "Source" form shall mean the preferred form for making modifications,
              including but not limited to software source code, documentation
              source, and configuration files.
        
              "Object" form shall mean any form resulting from mechanical
              transformation or translation of a Source form, including but
              not limited to compiled object code, generated documentation,
              and conversions to other media types.
        
              "Work" shall mean the work of authorship, whether in Source or
              Object form, made available under the License, as indicated by a
              copyright notice that is included in or attached to the work
              (an example is provided in the Appendix below).
        
              "Derivative Works" shall mean any work, whether in Source or Object
              form, that is based on (or derived from) the Work and for which the
              editorial revisions, annotations, elaborations, or other modifications
              represent, as a whole, an original work of authorship. For the purposes
              of this License, Derivative Works shall not include works that remain
              separable from, or merely link (or bind by name) to the interfaces of,
              the Work and Derivative Works thereof.
        
              "Contribution" shall mean any work of authorship, including
              the original version of the Work and any modifications or additions
              to that Work or Derivative Works thereof, that is intentionally
              submitted to Licensor for inclusion in the Work by the copyright owner
              or by an individual or Legal Entity authorized to submit on behalf of
              the copyright owner. For the purposes of this definition, "submitted"
              means any form of electronic, verbal, or written communication sent
              to the Licensor or its representatives, including but not limited to
              communication on electronic mailing lists, source code control systems,
              and issue tracking systems that are managed by, or on behalf of, the
              Licensor for the purpose of discussing and improving the Work, but
              excluding communication that is conspicuously marked or otherwise
              designated in writing by the copyright owner as "Not a Contribution."
        
              "Contributor" shall mean Licensor and any individual or Legal Entity
              on behalf of whom a Contribution has been received by Licensor and
              subsequently incorporated within the Work.
        
           2. Grant of Copyright License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              copyright license to reproduce, prepare Derivative Works of,
              publicly display, publicly perform, sublicense, and distribute the
              Work and such Derivative Works in Source or Object form.
        
           3. Grant of Patent License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              (except as stated in this section) patent license to make, have made,
              use, offer to sell, sell, import, and otherwise transfer the Work,
              where such license applies only to those patent claims licensable
              by such Contributor that are necessarily infringed by their
              Contribution(s) alone or by combination of their Contribution(s)
              with the Work to which such Contribution(s) was submitted. If You
              institute patent litigation against any entity (including a
              cross-claim or counterclaim in a lawsuit) alleging that the Work
              or a Contribution incorporated within the Work constitutes direct
              or contributory patent infringement, then any patent licenses
              granted to You under this License for that Work shall terminate
              as of the date such litigation is filed.
        
           4. Redistribution. You may reproduce and distribute copies of the
              Work or Derivative Works thereof in any medium, with or without
              modifications, and in Source or Object form, provided that You
              meet the following conditions:
        
              (a) You must give any other recipients of the Work or
                  Derivative Works a copy of this License; and
        
              (b) You must cause any modified files to carry prominent notices
                  stating that You changed the files; and
        
              (c) You must retain, in the Source form of any Derivative Works
                  that You distribute, all copyright, patent, trademark, and
                  attribution notices from the Source form of the Work,
                  excluding those notices that do not pertain to any part of
                  the Derivative Works; and
        
              (d) If the Work includes a "NOTICE" text file as part of its
                  distribution, then any Derivative Works that You distribute must
                  include a readable copy of the attribution notices contained
                  within such NOTICE file, excluding those notices that do not
                  pertain to any part of the Derivative Works, in at least one
                  of the following places: within a NOTICE text file distributed
                  as part of the Derivative Works; within the Source form or
                  documentation, if provided along with the Derivative Works; or,
                  within a display generated by the Derivative Works, if and
                  wherever such third-party notices normally appear. The contents
                  of the NOTICE file are for informational purposes only and
                  do not modify the License. You may add Your own attribution
                  notices within Derivative Works that You distribute, alongside
                  or as an addendum to the NOTICE text from the Work, provided
                  that such additional attribution notices cannot be construed
                  as modifying the License.
        
              You may add Your own copyright statement to Your modifications and
              may provide additional or different license terms and conditions
              for use, reproduction, or distribution of Your modifications, or
              for any such Derivative Works as a whole, provided Your use,
              reproduction, and distribution of the Work otherwise complies with
              the conditions stated in this License.
        
           5. Submission of Contributions. Unless You explicitly state otherwise,
              any Contribution intentionally submitted for inclusion in the Work
              by You to the Licensor shall be under the terms and conditions of
              this License, without any additional terms or conditions.
              Notwithstanding the above, nothing herein shall supersede or modify
              the terms of any separate license agreement you may have executed
              with Licensor regarding such Contributions.
        
           6. Trademarks. This License does not grant permission to use the trade
              names, trademarks, service marks, or product names of the Licensor,
              except as required for reasonable and customary use in describing the
              origin of the Work and reproducing the content of the NOTICE file.
        
           7. Disclaimer of Warranty. Unless required by applicable law or
              agreed to in writing, Licensor provides the Work (and each
              Contributor provides its Contributions) on an "AS IS" BASIS,
              WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
              implied, including, without limitation, any warranties or conditions
              of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
              PARTICULAR PURPOSE. You are solely responsible for determining the
              appropriateness of using or redistributing the Work and assume any
              risks associated with Your exercise of permissions under this License.
        
           8. Limitation of Liability. In no event and under no legal theory,
              whether in tort (including negligence), contract, or otherwise,
              unless required by applicable law (such as deliberate and grossly
              negligent acts) or agreed to in writing, shall any Contributor be
              liable to You for damages, including any direct, indirect, special,
              incidental, or consequential damages of any character arising as a
              result of this License or out of the use or inability to use the
              Work (including but not limited to damages for loss of goodwill,
              work stoppage, computer failure or malfunction, or any and all
              other commercial damages or losses), even if such Contributor
              has been advised of the possibility of such damages.
        
           9. Accepting Warranty or Additional Liability. While redistributing
              the Work or Derivative Works thereof, You may choose to offer,
              and charge a fee for, acceptance of support, warranty, indemnity,
              or other liability obligations and/or rights consistent with this
              License. However, in accepting such obligations, You may act only
              on Your own behalf and on Your sole responsibility, not on behalf
              of any other Contributor, and only if You agree to indemnify,
              defend, and hold each Contributor harmless for any liability
              incurred by, or claims asserted against, such Contributor by reason
              of your accepting any such warranty or additional liability.
        
           END OF TERMS AND CONDITIONS
        
           APPENDIX: How to apply the Apache License to your work.
        
              To apply the Apache License to your work, attach the following
              boilerplate notice, with the fields enclosed by brackets "[]"
              replaced with your own identifying information. (Don't include
              the brackets!)  The text should be enclosed in the appropriate
              comment syntax for the file format. We also recommend that a
              file or class name and description of purpose be included on the
              same "printed page" as the copyright notice for easier
              identification within third-party archives.
        
           Copyright [yyyy] [name of copyright owner]
        
           Licensed under the Apache License, Version 2.0 (the "License");
           you may not use this file except in compliance with the License.
           You may obtain a copy of the License at
        
               http://www.apache.org/licenses/LICENSE-2.0
        
           Unless required by applicable law or agreed to in writing, software
           distributed under the License is distributed on an "AS IS" BASIS,
           WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
           See the License for the specific language governing permissions and
           limitations under the License.
        
Keywords: jax,neural-networks,deep-learning,reinforcement learning,meta reinforcement learning,gridworld,minigrid,xland
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Natural Language :: English
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: License :: OSI Approved :: Apache Software License
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: jax>=0.4.13
Requires-Dist: jaxlib>=0.4.13
Requires-Dist: flax>=0.7.0
Requires-Dist: rich>=13.4.2
Provides-Extra: dev
Requires-Dist: ruff>=0.1.6; extra == "dev"
Requires-Dist: pre-commit>=3.3.3; extra == "dev"
Provides-Extra: baselines
Requires-Dist: matplotlib>=3.7.2; extra == "baselines"
Requires-Dist: imageio>=2.31.2; extra == "baselines"
Requires-Dist: imageio-ffmpeg>=0.4.9; extra == "baselines"
Requires-Dist: wandb>=0.15.10; extra == "baselines"
Requires-Dist: pyrallis>=0.3.1; extra == "baselines"
Requires-Dist: distrax>=0.1.4; extra == "baselines"
Requires-Dist: optax>=0.1.5; extra == "baselines"
Requires-Dist: orbax>=0.1.9; extra == "baselines"

<h1 align="center">XLand-MiniGrid</h1>

<p align="center">
    <a href="https://pypi.python.org/pypi/xminigrid">
        <img src="https://img.shields.io/pypi/pyversions/xminigrid.svg"/>
    </a>
    <a href="https://badge.fury.io/py/xminigrid">
        <img src="https://badge.fury.io/py/xminigrid.svg"/>
    </a>
    <a href="https://badge.fury.io/py/xminigrid">
        <img src="https://img.shields.io/pypi/dm/xminigrid?color=yellow&label=Downloads"/>
    </a>
    <a href="https://github.com/corl-team/xland-minigrid/main/LICENSE">
        <img src="https://img.shields.io/badge/license-Apache_2.0-blue"/>
    </a>
    <a href="https://github.com/astral-sh/ruff">
        <img src="https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json"/>
    </a>
    <a href="https://twitter.com/vladkurenkov/status/1731709425524543550">
        <img src="https://badgen.net/badge/icon/twitter?icon=twitter&label"/>
    </a>
    <a target="_blank" href="https://colab.research.google.com/github/corl-team/xland-minigrid/blob/main/examples/walkthrough.ipynb">
      <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/>
    </a>
</p>

![img](figures/readme-main-img.png)

# Meta-Reinforcement Learning in JAX

> 🥳 **XLand-MiniGrid was [accepted](https://openreview.net/forum?id=xALDC4aHGz) to [Intrinsically Motivated Open-ended Learning](https://imol-workshop.github.io) workshop at NeurIPS 2023.** We look forward to seeing everyone at the poster session! 

**XLand-MiniGrid** is a suite of tools, grid-world environments and benchmarks for meta-reinforcement learning research inspired by 
the diversity and depth of [XLand](https://deepmind.google/discover/blog/generally-capable-agents-emerge-from-open-ended-play/) 
and the simplicity and minimalism of [MiniGrid](https://github.com/Farama-Foundation/MiniGrid). Despite the similarities, 
XLand-MiniGrid is written in JAX from scratch and designed to be highly scalable, democratizing large-scale experimentation 
with limited resources. Ever wanted to reproduce a [DeepMind AdA](https://sites.google.com/view/adaptive-agent/) agent? Now you can and not in months, but days!


### Features

- 🔮 System of rules and goals that can be combined in arbitrary ways to produce
diverse task distributions
- 🔧 Simple to extend and modify, comes with example environments ported from the original
[MiniGrid](https://github.com/Farama-Foundation/MiniGrid)
- 🪄 Fully compatible with all JAX transformations, can run on CPU, GPU and TPU
- 📈 Easily scales to $2^{16}$ parallel environments and millions of steps per second on a single GPU
- 🔥 Multi-GPU PPO baselines in the [PureJaxRL](https://github.com/luchris429/purejaxrl) style, which can achieve **1 trillion** environment steps under two days 

How cool is that? For more details, take a look at the [technical paper (soon)]() or
[examples](examples), which will walk you through the basics and training your own adaptive agents in minutes!

![img](figures/times_minigrid.jpg)
TODO: update this with the latest version of the codebase...

## Installation 🎁

> ⚠️ XLand-MiniGrid is currently in alpha stage, so expect breaking changes! ⚠️

The latest release of XLand-MiniGrid can be installed directly from PyPI:

```commandline
pip install xminigrid
# or, from github directly
pip install "xminigrid @ git+https://github.com/corl-team/xland-minigrid.git"
```

Alternatively, if you want to install the latest development version from the GitHub and run provided algorithms or scripts,
install the source as follows:
```commandline
git clone git@github.com:corl-team/xland-minigrid.git
cd xland-minigrid
# additional dependencies for baselines
pip install -e ".[dev,benchmark]"
```
Note that the installation of JAX may differ depending on your hardware accelerator! 
We advise users to explicitly install the correct JAX version (see the [official installation guide](https://github.com/google/jax#installation)).

## Basic Usage 🕹️

Most users who are familiar with other popular JAX-based environments 
(such as [gymnax](https://github.com/RobertTLange/gymnax) or [jumnaji](https://github.com/instadeepai/jumanji)), 
will find that the interface is very similar.
On the high level, current API combines [dm_env](https://github.com/google-deepmind/dm_env) and gymnax interfaces.

```python
import jax
import xminigrid

key = jax.random.PRNGKey(0)
reset_key, ruleset_key = jax.random.split(key)

# to list available benchmarks: xminigrid.registered_benchmarks()
benchmark = xminigrid.load_benchmark(name="trivial-1m")
# choosing ruleset, see section on rules and goals
ruleset = benchmark.sample_ruleset(ruleset_key)

# to list available environments: xminigrid.registered_environments()
env, env_params = xminigrid.make("XLand-MiniGrid-R9-25x25")
env_params = env_params.replace(ruleset=ruleset)

# fully jit-compatible step and reset methods
timestep = jax.jit(env.reset)(env_params, reset_key)
timestep = jax.jit(env.step)(env_params, timestep, action=0)

# optionally render the state
env.render(env_params, timestep)
```
Similar to gymnasium or jumanji, users can register new environment 
variations with `register` for convenient further usage with `make`. 
`timestep` is a dataclass containing `step_type`, `reward`, `discount`, `observation`, as well as the internal environment `state`.

For a bit more advanced introduction see provided [walkthrough notebook](examples/walkthrough.ipynb).

### On environment interface

Currently, there are a lot of new JAX-based environments appearing, each offering its own variant of API. Initially, we tried to reuse Jumaji, but it turned out 
that its design [is not suitable for meta learning](https://github.com/instadeepai/jumanji/issues/212). The Gymnax design appeared to be more appropriate, but unfortunately it is not actively supported and
often departs from the idea that parameters should only be contained in `env_params`. Furthermore, splitting 
`timestep` into multiple entities seems suboptimal to us, as it complicates many things, such as envpool or dm_env 
style auto reset, where the reset occurs on the next step (we need access to done of previous step).

Therefore, we decided that we would make a minimal interface that would cover just our needs without the 
goal of making it generic. The core of our library is interface independent, and we plan 
to switch to the new one when/if a better design becomes available
(e.g. when stable Gymnasium [FuncEnv](https://gymnasium.farama.org/main/api/functional/) is released).

## Rules and Goals 🔮

<img src="figures/ruleset-example.jpg" align="right" width="55%" style="margin:15px;">

In XLand-MiniGrid, the system of rules and goals is the cornerstone of the 
emergent complexity and diversity. In the original MiniGrid 
some environments have dynamic goals, but the dynamics are never changed. 
To train and evaluate highly adaptive agents, we need to be able to change 
the dynamics in non-trivial ways. 

**Rules** are the functions that can change the environment state in some deterministic 
way according to the given conditions. **Goals** are similar to rules, except they do 
not change the state, they only test conditions. Every task should be described with a goal, rules and initial objects. We call these **rulesets**. 
Currently, we support only one goal per task. 

To illustrate, we provide visualization for specific ruleset. To solve this task, agent should take blue pyramid and put it near the purple square to transform both 
objects into red circle. To complete the goal, red circle should be placed near
green circle. However, placing purple square near yellow circle will make it unsolvable in this trial. Initial objects positions will be randomized on each reset. 

For more advanced introduction, see corresponding section in the provided [walkthrough notebook](examples/walkthrough.ipynb).
<br clear="right"/>

## Benchmarks 🎲 

While composing rules and goals by hand is flexible, it can quickly become cumbersome. 
Besides, it's hard to express efficiently in a JAX-compatible way due to the high number of heterogeneous computations 

To avoid significant overhead during training and facilitate reliable comparisons between agents, 
we pre-sampled several benchmarks with up to **five million unique tasks**, following the procedure used to train DeepMind 
AdA agent from the original XLand. These benchmarks differ in the generation configs, producing distributions with
varying levels of diversity and average difficulty of the tasks. They can be used for different purposes, for example
the `trivial-1m` benchmark can be used to debug your agents, allowing very quick iterations. However, we would caution 
against treating benchmarks as a progression from simple to complex. They are just different 🤷.

Pre-sampled benchmarks are hosted on [HuggingFace](https://huggingface.co/datasets/Howuhh/xland_minigrid/tree/main) and will be downloaded and cached on the first use:

```python
import jax.random
import xminigrid
from xminigrid.benchmarks import Benchmark

# downloading to path specified by XLAND_MINIGRID_DATA,
# ~/.xland_minigrid by default
benchmark: Benchmark = xminigrid.load_benchmark(name="trivial-1m")
# reusing cached on the second use
benchmark: Benchmark = xminigrid.load_benchmark(name="trivial-1m")

# users can sample or get specific rulesets
benchmark.sample_ruleset(jax.random.PRNGKey(0))
benchmark.get_ruleset(ruleset_id=benchmark.num_rulesets() - 1)

# or split them for train & test
train, test = benchmark.shuffle(key=jax.random.PRNGKey(0)).split(prop=0.8)
```

We also provide the [script](scripts/ruleset_generator.py) used to generate these benchmarks. Users can use it for their own purposes:
```commandline
python scripts/ruleset_generator.py --help
```

In depth description of all available benchmarks is provided [here (soon)]().

**P.S.** Be aware, that benchmarks can change, as we are currently testing and balancing them!

## Environments 🌍

We provide environments from two domains. `XLand` is our main focus for meta-learning. For this domain we provide single
environment and numerous registered variants with different grid layouts and sizes. All of them can be combined
with arbitrary rulesets. 

To demonstrate the generality of our library we also port majority of 
non-language based tasks from original `MiniGrid`. Similarly, some environments come with multiple registered variants. 
However, we have no current plans to actively develop and support them (but that may change).

| Name | Domain  | Visualization                                            | Goal                                                                         |
|------|---------|----------------------------------------------------------|------------------------------------------------------------------------------|
|   `XLand-MiniGrid` | XLand   | <img src="figures/xland.png" width="90px">               | specified by the provided ruleset                                            |
|   `MiniGrid-Empty`   | MiniGrid | <img src="figures/empty.png" width="90px">               | go to the green goal                                                         |
|   `MiniGrid-EmptyRandom`   | MiniGrid | <img src="figures/empty_random.png" width="90px">        | go the green goal from different starting positions                          |
|   `MiniGrid-FourRooms`   | MiniGrid | <img src="figures/fourrooms.png" width="90px">           | go the green goal, but goal and starting positions are randomized            |
|   `MiniGrid-LockedRoom`   | MiniGrid | <img src="figures/lockedroom.png" width="90px">          | find the key to unlock the door, go to the green goal                        |
|   `MiniGrid-Memory`   | MiniGrid | <img src="figures/memory.png" width="90px">              | remember the initial object and choose it at the end of the corridor         |
|   `MiniGrid-Playground`   | MiniGrid | <img src="figures/playground.png" width="90px">          | goal is not specified                                                        |
|   `MiniGrid-Unlock`   | MiniGrid | <img src="figures/unlock.png" width="90px">              | unlock the door with the key                                                 |
|   `MiniGrid-UnlockPickUp`   | MiniGrid | <img src="figures/unlockpickup.png" width="90px">        | unlock the door and pick up the object in another room                       |
|   `MiniGrid-BlockedUnlockPickUp`   | MiniGrid | <img src="figures/blockedunlockpickup.png" width="90px"> | unlock the door blocked by the object and pick up the object in another room |
|   `MiniGrid-DoorKey`   | MiniGrid | <img src="figures/doorkey.png" width="90px">             | unlock the door and go to the green goal                                     |

Users can get all registered environments with `xminigrid.registered_environments()`. We also provide manual control to easily explore the environments:
```commandline
python -m xminigrid.manual_control --env-id="MiniGrid-Empty-8x8"
```

## Baselines 🚀

In addition to the environments, we provide high-quality *almost* single-file 
implementations of recurrent PPO baselines in the style of [PureJaxRL](https://github.com/luchris429/purejaxrl). With the help of magical `jax.pmap` transformation 
they can scale to multiple accelerators, achieving impressive FPS of millions during training. 

Agents can be trained from the terminal and default arguments can be overwritten from the command line or from the yaml config:
```commandline
# for meta learning
python training/train_meta_task.py \
    --config-path='some-path/config.yaml' \
    --env_id='XLand-MiniGrid-R1-9x9'

# for minigrid envs
python training/train_singe_task.py \
    --config-path='some-path/config.yaml' \ 
    --env_id='XLand-MiniGrid-R1-9x9'
```
For the source code and hyperparameters available see [/training](training) or run `python training/train_meta_task.py --help`. 
Furthermore, we provide standalone implementations that can be trained in Colab:
[xland](examples/train_meta_standalone.ipynb),
[minigrid](examples/train_single_standalone.ipynb). 

**P.S.** Do not expect that provided baselines will solve the hardest environments or benchmarks 
available. How much fun would that be 🤔? However, we hope that they will 
help to get started quickly!

## Roadmap 🗓️

With the initial release of XLand-MiniGrid, things are just getting started. There is a long way to go in 
terms of polishing the code, adding new features, and improving the overall user experience. What we
currently plan to improve in forthcoming releases:
1. Tweaks to the benchmark generation, time-limits
2. Documentation (in code and as standalone site)
3. Full type hints coverage, type checking
4. Tests 
5. More examples and tutorials

After that we will start thinking on new major features, environments and bechmarks. 
However, we should perfect the core before that.

## Contributing 🔨

We welcome anyone interested in helping out! Please take a look at our [contribution guide](CONTRIBUTING.md) 
for further instructions and open an issue if something is not clear.

## See Also 🔎

A lot of other work is going in a similar direction, transforming RL through JAX. Many of them have inspired us, 
and we encourage users to check them out as well.

- [Brax](https://github.com/google/brax) - fully differentiable physics engine used for research and development of robotics.
- [Gymnax](https://github.com/RobertTLange/gymnax) - implements classic environments including classic control, bsuite, MinAtar and simplistic meta learning tasks.
- [Jumanji](https://github.com/instadeepai/jumanji) - a diverse set of environments ranging from simple games to NP-hard combinatorial problems.
- [Pgx](https://github.com/sotetsuk/pgx) - JAX implementations of classic board games, such as Chess, Go and Shogi.
- [JaxMARL](https://github.com/flairox/jaxmarl) - multi-agent RL in JAX with wide range of commonly used environments.

Let's build together!

## Citation 🙏

```bibtex
@inproceedings{
    nikulin2023xlandminigrid,
    title={{XL}and-MiniGrid: Scalable Meta-Reinforcement Learning Environments in {JAX}},
    author={Alexander Nikulin and Vladislav Kurenkov and Ilya Zisman and Viacheslav Sinii and Artem Agarkov and Sergey Kolesnikov},
    booktitle={Intrinsically-Motivated and Open-Ended Learning Workshop, NeurIPS2023},
    year={2023},
    url={https://openreview.net/forum?id=xALDC4aHGz}
}
```
