MarlGroupMapType

torchrl.envs.MarlGroupMapType(value, names=None, *, module=None, qualname=None, type=None, start=1)

Marl Group Map 类型。

作为 torchrl 多智能体的一个特性，你可以控制环境中智能体的分组。你可以将智能体组合在一起（堆叠它们的张量），以便在通过相同的神经网络时利用向量化。你可以将智能体分成不同的组，在这些组中，它们是异构的，或者应该由不同的神经网络处理。要分组，你只需要在环境构建时传递一个 group_map。

否则，你可以从此类中选择预制的分组策略之一。

• 使用 group_map=MarlGroupMapType.ALL_IN_ONE_GROUP 和智能体 ["agent_0", "agent_1", "agent_2", "agent_3"]，来自和去往你的环境的 tensordict 将如下所示

  >>> print(env.rand_action(env.reset()))
  TensorDict(
      fields={
          agents: TensorDict(
              fields={
                  action: Tensor(shape=torch.Size([4, 9]), device=cpu, dtype=torch.int64, is_shared=False),
                  done: Tensor(shape=torch.Size([4, 1]), device=cpu, dtype=torch.bool, is_shared=False),
                  observation: Tensor(shape=torch.Size([4, 3, 3, 2]), device=cpu, dtype=torch.int8, is_shared=False)},
              batch_size=torch.Size([4]))},
      batch_size=torch.Size([]))
  >>> print(env.group_map)
  {"agents": ["agent_0", "agent_1", "agent_2", "agent_3]}
  

• 使用 group_map=MarlGroupMapType.ONE_GROUP_PER_AGENT 和智能体 ["agent_0", "agent_1", "agent_2", "agent_3"]，来自和去往你的环境的 tensordict 将如下所示

  >>> print(env.rand_action(env.reset()))
  TensorDict(
      fields={
          agent_0: TensorDict(
              fields={
                  action: Tensor(shape=torch.Size([9]), device=cpu, dtype=torch.int64, is_shared=False),
                  done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False),
                  observation: Tensor(shape=torch.Size([3, 3, 2]), device=cpu, dtype=torch.int8, is_shared=False)},
              batch_size=torch.Size([]))},
          agent_1: TensorDict(
              fields={
                  action: Tensor(shape=torch.Size([9]), device=cpu, dtype=torch.int64, is_shared=False),
                  done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False),
                  observation: Tensor(shape=torch.Size([3, 3, 2]), device=cpu, dtype=torch.int8, is_shared=False)},
              batch_size=torch.Size([]))},
          agent_2: TensorDict(
              fields={
                  action: Tensor(shape=torch.Size([9]), device=cpu, dtype=torch.int64, is_shared=False),
                  done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False),
                  observation: Tensor(shape=torch.Size([3, 3, 2]), device=cpu, dtype=torch.int8, is_shared=False)},
              batch_size=torch.Size([]))},
          agent_3: TensorDict(
              fields={
                  action: Tensor(shape=torch.Size([9]), device=cpu, dtype=torch.int64, is_shared=False),
                  done: Tensor(shape=torch.Size([1]), device=cpu, dtype=torch.bool, is_shared=False),
                  observation: Tensor(shape=torch.Size([3, 3, 2]), device=cpu, dtype=torch.int8, is_shared=False)},
              batch_size=torch.Size([]))},
      batch_size=torch.Size([]))
  >>> print(env.group_map)
  {"agent_0": ["agent_0"], "agent_1": ["agent_1"], "agent_2": ["agent_2"], "agent_3": ["agent_3"]}