Metadata-Version: 2.1
Name: dexterity
Version: 0.0.12
Summary: Software and tasks for dexterous multi-fingered hand manipulation, powered by MuJoCo
Home-page: https://github.com/kevinzakka/dexterity
Author: Kevin Zakka
Author-email: kevinarmandzakka@gmail.com
Maintainer: Kevin Zakka
Maintainer-email: kevinarmandzakka@gmail.com
License: MIT
Platform: UNKNOWN
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: MIT License
Classifier: Natural Language :: English
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.7
Description-Content-Type: text/markdown
Provides-Extra: testing
Provides-Extra: examples
Provides-Extra: dev
License-File: LICENSE

# The MuJoCo Dexterity Suite (alpha-release)

[![PyPI Python Version][pypi-versions-badge]][pypi]
[![PyPI version][pypi-badge]][pypi]
[![dexterity-tests][tests-badge]][tests]

[pypi-versions-badge]: https://img.shields.io/pypi/pyversions/dexterity
[pypi-badge]: https://badge.fury.io/py/dexterity.svg
[pypi]: https://pypi.org/project/dexterity/
[tests-badge]: https://github.com/kevinzakka/dexterity/actions/workflows/build.yml/badge.svg
[tests]: https://github.com/kevinzakka/dexterity/actions/workflows/build.yml

Software and tasks for dexterous multi-fingered hand manipulation, powered by [MuJoCo](https://mujoco.org/).

<p float="left">
  <img src="https://raw.githubusercontent.com/kevinzakka/dexterity/main/assets/reach.png" height="200">
  <img src="https://raw.githubusercontent.com/kevinzakka/dexterity/main/assets/cube.png" height="200">
</p>

`dexterity` builds on [dm_control](https://github.com/deepmind/dm_control) and provides a collection of modular components that can be used to define rich Reinforcement Learning environments for dexterous manipulation. It also comes with a set of standardized tasks that can serve as a performance benchmark for the research community.

An introductory tutorial is available as a Colab notebook: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/kevinzakka/dexterity/blob/main/tutorial.ipynb)

## Installation

### PyPI (Recommended)

The recommended way to install this package is via [PyPI](https://pypi.org/project/dexterity/):

```bash
pip install dexterity
```

### Source

We provide a [Miniconda](https://docs.conda.io/en/latest/miniconda.html) environment with Python 3.8 for development. To create it and install dependencies, run the following steps:

```bash
git clone https://github.com/kevinzakka/dexterity
cd dexterity
conda env create -f environment.yml  # Creates a dexterity env.
conda activate dexterity
pip install .
```

## Overview

The MuJoCo `dexterity` suite is composed of the following core components:

* [`models`](dexterity/models/): MuJoCo models for dexterous hands and [PyMJCF](https://github.com/deepmind/dm_control/blob/main/dm_control/mjcf/README.md) classes for dynamically customizing them.
* [`inverse_kinematics`](dexterity/inverse_kinematics/): Inverse kinematics solver for multi-fingered hands.
* [`effectors`](dexterity/effectors/): Interfaces for controlling hands and defining action spaces.

These components in conjunction with `dm_control` allow you to define and customize rich environments for reinforcement learning.

`dexterity` also comes pre-packaged with a suite of benchmark RL environments. Our hope is to grow it over time with crowd-sourced contributions from the research community. For an overview of the current tasks, see the [task library](dexterity/manipulation/README.md).

## Acknowledgements

A large part of the design and implementation of `dexterity` is inspired by the [MoMa](https://github.com/deepmind/dm_robotics/tree/main/py/moma) library in [dm_robotics](https://github.com/deepmind/dm_robotics/).


