Metadata-Version: 2.1
Name: drlgeb
Version: 0.0.1
Summary: A sample but effective drl lib.
Home-page: https://github.com/mikuh/drlgeb
Author: geb
Author-email: 853934146@qq.com
License: UNKNOWN
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.7
Description-Content-Type: text/markdown
Requires-Dist: numpy
Requires-Dist: tensorflow
Requires-Dist: gym
Requires-Dist: gym[atari]

# drlgeb
A sample DRL lib use tf.keras and gym.
Just writing ...

Now just implement the batch a3c, other algorithms will be updated in the future.

## Requirement
- Linux/windows/osx
- python3.8
- tensorflow2.4
- gym[atari]
- numpy

## Installation
```shell script
pip install drlgeb
```

## Get start

#### A3C
```python
from drlgeb.ac import A3C



env_id = "SpaceInvaders-v0"
agent = A3C(env_id=env_id)

# train
agent.learn()

# test
model_path = "..."
agent.play(episodes=5, model_path=model_path)

```


