|
|
|
|
|
|
|
the `gym-unity` package to perform reinforcement learning within Unity. `mlagents_envs` can be |
|
|
|
used independently of `mlagents` for Python communication. |
|
|
|
|
|
|
|
The `mlagents_envs` Python package contains one sub package: |
|
|
|
|
|
|
|
* `mlagents_envs`: A low level API which allows you to interact directly with a |
|
|
|
Unity Environment. See |
|
|
|
[here](../docs/Python-API.md) |
|
|
|
for more information on using this package. |
|
|
|
The `mlagents_envs` Python package contains a low level API which allows you to interact |
|
|
|
directly with a Unity Environment. See [here](../docs/Python-API.md) for more information |
|
|
|
on using this API. |
|
|
|
|
|
|
|
## Installation |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
For more detailed documentation, check out the |
|
|
|
[ML-Agents Toolkit documentation.](../docs/Readme.md) |
|
|
|
|
|
|
|
## Limitations |
|
|
|
- `mlagents_envs` uses localhost ports to exchange data between Unity and Python. As such, |
|
|
|
multiple instances can have their ports collide leading to errors. Make sure to use a |
|
|
|
different port if you are using multiple instances of `UnityEnvironment`. |
|
|
|
- Communication between Unity and the Python `UnityEnvironment` is not secure. |
|
|
|
- On Linux, ports are not released immediately after the communication closes. As such, you |
|
|
|
cannot reuse ports right after closing a `UnityEnvironment`. |
GitHub