https://github.com/danijar/embodied
Fast reinforcement learning research
https://github.com/danijar/embodied
artificial-intelligence deep-learning reinforcement-learning research science tools
Last synced: 11 months ago
JSON representation
Fast reinforcement learning research
- Host: GitHub
- URL: https://github.com/danijar/embodied
- Owner: danijar
- License: mit
- Created: 2016-01-24T02:01:09.000Z (over 10 years ago)
- Default Branch: main
- Last Pushed: 2024-12-07T21:47:54.000Z (over 1 year ago)
- Last Synced: 2025-07-09T08:04:29.080Z (12 months ago)
- Topics: artificial-intelligence, deep-learning, reinforcement-learning, research, science, tools
- Language: Python
- Homepage: https://embodied.readthedocs.org
- Size: 741 KB
- Stars: 61
- Watchers: 9
- Forks: 15
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
[](https://pypi.python.org/pypi/embodied/#history)
# Embodied
Fast reinforcement learning research.
## Overview
The goal of Embodied is to empower researchers to quickly implement new agents
at scale. Embodied achieves this by specifying an interface both for
environments and agents, allowing users to mix and match agents, envs, and
evaluation protocols. Embodied provides common building blocks that users are
encouraged to fork when more control is needed. The only dependency is Numpy
and agents can be implemented in any framework.
## Packages
```
embodied/
core/ # Config, logging, checkpointing, simulation, wrappers
run/ # Evaluation protocols that combine agents and environments
envs/ # Environment suites such as Gym, Atari, DMC, Crafter
agents/ # Agent implementations
```
## Agent API
```python
class Agent:
__init__(obs_space, act_space, config)
policy(obs, carry, mode='train') -> act, carry
train(data, carry) -> metrics, carry
report(data, carry) -> metrics, carry
init_policy(batch_size) -> carry
init_train(batch_size) -> carry
init_report(batch_size) -> carry
dataset(generator) -> generator
```
## Env API
```python
class Env:
__len__() -> int
@obs_space -> dict of spaces
@act_space -> dict of spaces
step(act) -> obs dict
render() -> array
close()
```