ModelBasedEnvBase

torchrl.envs.ModelBasedEnvBase(*args, **kwargs)

模型驅動強化學習 (Model Based RL) SOTA 實現的基礎環境。

對 MBRL 演算法的模型進行包裝。旨在為世界模型（包括但不限於觀測、獎勵、完成狀態和安全約束模型）提供一個環境框架，並作為經典環境執行。

這是一個其他環境的基類，不應直接使用。

示例

>>> import torch
>>> from tensordict import TensorDict
>>> from torchrl.data import Composite, Unbounded
>>> class MyMBEnv(ModelBasedEnvBase):
...     def __init__(self, world_model, device="cpu", dtype=None, batch_size=None):
...         super().__init__(world_model, device=device, dtype=dtype, batch_size=batch_size)
...         self.observation_spec = Composite(
...             hidden_observation=Unbounded((4,))
...         )
...         self.state_spec = Composite(
...             hidden_observation=Unbounded((4,)),
...         )
...         self.action_spec = Unbounded((1,))
...         self.reward_spec = Unbounded((1,))
...
...     def _reset(self, tensordict: TensorDict) -> TensorDict:
...         tensordict = TensorDict(
...             batch_size=self.batch_size,
...             device=self.device,
...         )
...         tensordict = tensordict.update(self.state_spec.rand())
...         tensordict = tensordict.update(self.observation_spec.rand())
...         return tensordict
>>> # This environment is used as follows:
>>> import torch.nn as nn
>>> from torchrl.modules import MLP, WorldModelWrapper
>>> world_model = WorldModelWrapper(
...     TensorDictModule(
...         MLP(out_features=4, activation_class=nn.ReLU, activate_last_layer=True, depth=0),
...         in_keys=["hidden_observation", "action"],
...         out_keys=["hidden_observation"],
...     ),
...     TensorDictModule(
...         nn.Linear(4, 1),
...         in_keys=["hidden_observation"],
...         out_keys=["reward"],
...     ),
... )
>>> env = MyMBEnv(world_model)
>>> tensordict = env.rollout(max_steps=10)
>>> print(tensordict)
TensorDict(
    fields={
        action: Tensor(torch.Size([10, 1]), dtype=torch.float32),
        done: Tensor(torch.Size([10, 1]), dtype=torch.bool),
        hidden_observation: Tensor(torch.Size([10, 4]), dtype=torch.float32),
        next: LazyStackedTensorDict(
            fields={
                hidden_observation: Tensor(torch.Size([10, 4]), dtype=torch.float32)},
            batch_size=torch.Size([10]),
            device=cpu,
            is_shared=False),
        reward: Tensor(torch.Size([10, 1]), dtype=torch.float32)},
    batch_size=torch.Size([10]),
    device=cpu,
    is_shared=False)

屬性

observation_spec (Composite): 觀測的取樣規範；action_spec (TensorSpec): 行動的取樣規範；reward_spec (TensorSpec): 獎勵的取樣規範；input_spec (Composite):輸入的取樣規範；batch_size (torch.Size): 環境使用的批處理大小。如果未設定，環境將接受所有批處理大小的 tensordicts。device (torch.device): 環境輸入和輸出期望所在的裝置

引數:

• world_model (nn.Module) – 生成世界狀態及其相應獎勵的模型；

• params (List[torch.Tensor], optional) – 世界模型的引數列表；

• buffers (List[torch.Tensor], optional) – 世界模型的緩衝區列表；

• device (torch.device, optional) – 環境輸入和輸出期望所在的裝置

• dtype (torch.dtype, optional) – 環境輸入和輸出的資料型別

• batch_size (torch.Size, optional) – 例項中包含的環境數量

• run_type_check (bool, optional) – 是否在環境的步進過程中執行型別檢查

torchrl.envs.step(TensorDict -> TensorDict)

環境中的步進

torchrl.envs.reset(TensorDict, optional -> TensorDict)

重置環境

torchrl.envs.set_seed(int -> int)

設定環境的種子

torchrl.envs.rand_step(TensorDict, optional -> TensorDict)

根據動作規範進行隨機步進

torchrl.envs.rollout(Callable, ... -> TensorDict)

使用給定的策略（如果未提供策略，則為隨機步進）在環境中執行滾動。