---
tags: 2021-UnityRL
---

# Unity Python API
> Credit by Chien-Feng

## Documentations
1. 路徑設定(pip)

Inside the ml-agents directory

    cd gym_unity/
    pip3 install -e .

    cd ml-agents-envs/
    pip3 install -e .

2. 實作細節

在實作上，如果要直接跟 editor 互動的話，要把 UnityEnvironment 中的 file_name 設為 None：

    UnityEnvironment(file_name=None)

UnityToGymWrapper 中的 uint8_visual 如果是 True 代表說吃到的影像範圍為 0-255，defualt 的話為 False。另外這邊和原本的 doc 有些出入，原本的 doc 上寫的是：

    UnityToGymWrapper(unity_env, 0, uint8_visual=True)

然而後來看了一下 code 發現這個 0 不知所云，可能是版本問題，總之要拿掉才跑得動：

    UnityToGymWrapper(unity_env, uint8_visual=True)

整體而言用法如：reset、step 這些都與與 gym 相同，有些更詳細的用法可能可以自己去 code 裡面翻翻看。

3. 範例

OpenAI baselines 的安裝：https://github.com/openai/baselines
Sample code with OpenAI baselines

    import gym

    from baselines import deepq
    from baselines import logger

    from mlagents_envs.environment import UnityEnvironment
    from gym_unity.envs import UnityToGymWrapper

    def main():
        unity_env = UnityEnvironment(file_name=None)
        env = UnityToGymWrapper(unity_env, uint8_visual=True)
        logger.configure('./logs') # Change to log in a different directory

        act = deepq.learn(
            env,
            "cnn", # conv_only is also a good choice for GridWorld
            lr=2.5e-4,
            total_timesteps=1000000,
            buffer_size=50000,
            exploration_fraction=0.05,
            exploration_final_eps=0.1,
            print_freq=20,
            train_freq=5,
            learning_starts=20000,
            target_network_update_freq=50,
            gamma=0.99,
            prioritized_replay=False,
            checkpoint_freq=1000,
            checkpoint_path='./logs', # Change to save model in a different directory
            dueling=True
        )
        print("Saving model to unity_model.pkl")
        act.save("unity_model.pkl")

    if __name__ == '__main__':
        main()

執行以上程式碼後會看到 Terminal 印出一行字：Start training by pressing the Play button in the Unity Editor，此時再按下 Editor 上的 Play 鍵就開始訓練了。

## Reference
1. UnityEnvironment: https://github.com/Unity-Technologies/ml-agents/blob/main/docs/Python-API.md
2. UnityToGymWrapper: https://github.com/Unity-Technologies/ml-agents/blob/main/gym-unity/README.md