UnityMLAgentsEnv¶
- torchrl.envs.UnityMLAgentsEnv(*args, **kwargs)[源代码]¶
Unity ML-Agents 环境包装器。
GitHub: https://github.com/Unity-Technologies/ml-agents
文档: https://unity-technologies.github.io/ml-agents/Python-LLAPI/
此类可以提供
mlagents_envs.environment.UnityEnvironment类提供的任何可选初始化参数。有关这些参数的列表，请参阅：https://unity-technologies.github.io/ml-agents/Python-LLAPI-Documentation/#__init__如果同时提供
file_name和registered_name，则会引发错误。如果既未提供
file_name也未提供registered_name，则环境设置将等待本地主机端口，用户必须执行 Unity ML-Agents 环境二进制文件才能连接到它。- 参数:
file_name (str, 可选) – 如果提供，则为 Unity 环境二进制文件的路径。默认为
None。registered_name (str, 可选) – 如果提供，则从默认 ML-Agents 注册表中加载 Unity 环境二进制文件。注册环境的列表位于
available_envs中。默认为None。
- 关键字参数:
device (torch.device, 可选) – 如果提供，则为数据要转换到的设备。默认为
None。batch_size (torch.Size, 可选) – 环境的批大小。默认为
torch.Size([])。allow_done_after_reset (bool, 可选) – 如果为
True，则在调用reset()后容许环境处于done状态。默认为False。group_map (MarlGroupMapType 或 Dict[str, List[str]]], 可选) – 如何在 tensordict 中对智能体进行分组以进行输入/输出。有关更多信息，请参阅
MarlGroupMapType。如果未指定，则根据 Unity 环境给出的组 ID 对智能体进行分组。默认为None。categorical_actions (bool, 可选) – 如果为
True，则分类规范将转换为 TorchRL 等效项 (torchrl.data.Categorical)，否则将使用 one-hot 编码 (torchrl.data.OneHot)。默认为False。
- 变量:
available_envs – 可用于构建的注册环境列表
示例
>>> from torchrl.envs import UnityMLAgentsEnv >>> env = UnityMLAgentsEnv(registered_name='3DBall') >>> td = env.reset() >>> td = env.step(td.update(env.full_action_spec.rand())) >>> td TensorDict( fields={ group_0: TensorDict( fields={ agent_0: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_10: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_11: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_1: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_2: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_3: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_4: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_5: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_6: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_7: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_8: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_9: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), continuous_action: Tensor(shape=torch.Size([2]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), next: TensorDict( fields={ group_0: TensorDict( fields={ agent_0: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_10: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_11: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_1: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_2: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_3: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_4: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_5: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_6: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_7: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_8: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False), agent_9: TensorDict( fields={ VectorSensor_size8: Tensor(shape=torch.Size([8]), device=cpu, dtype=torch.float32, is_shared=False), done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), group_reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), reward: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.float32, is_shared=False), terminated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False), truncated: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False)}, batch_size=torch.Size([]), device=None, is_shared=False)
