BatchSizeTransform¶

class torchrl.envs.transforms.BatchSizeTransform(*, batch_size: torch.Size | None = None, reshape_fn: Callable[[TensorDictBase], TensorDictBase] | None = None, reset_func: Callable[[TensorDictBase, TensorDictBase], TensorDictBase] | None = None, env_kwarg: bool = False)[源代码]¶

用于修改环境批大小的转换。

此转换有两种截然不同的用法：它可以用于设置非批次锁定（例如无状态）环境的批大小，以启用使用数据收集器进行数据收集。它还可以用于修改环境的批大小（例如压缩、解压缩或重塑）。

此转换会修改环境批大小以匹配提供的批大小。它期望父环境的批大小可扩展到提供的批大小。

关键字参数：

batch_size (torch.Size 或 等效项，可选) – 环境的新批大小。与 reshape_fn 互斥。
reshape_fn (可调用对象，可选) –
用于修改环境批大小的可调用对象。与 batch_size 互斥。

注意

目前，支持涉及 reshape、flatten、unflatten、squeeze 和 unsqueeze 的转换。如果需要其他重塑操作，请在 TorchRL github 上提交功能请求。
reset_func (可调用对象，可选) – 生成重置 tensordict 的函数。签名必须匹配 Callable[[TensorDictBase, TensorDictBase], TensorDictBase]，其中第一个输入参数是在调用 reset() 时传递给环境的可选 tensordict，第二个参数是 TransformedEnv.base_env.reset 的输出。如果 env_kwarg=True，它也可以支持可选的 env 关键字参数。
env_kwarg (布尔值，可选) – 如果为 True，则 reset_func 必须支持 env 关键字参数。默认为 False。传递的 env 将是伴随其转换的 env。

示例

>>> # Changing the batch-size with a function
>>> from torchrl.envs import GymEnv
>>> base_env = GymEnv("CartPole-v1")
>>> env = TransformedEnv(base_env, BatchSizeTransform(reshape_fn=lambda data: data.reshape(1, 1)))
>>> env.rollout(4)
>>> # Setting the shape of a stateless environment
>>> class MyEnv(EnvBase):
...     batch_locked = False
...     def __init__(self):
...         super().__init__()
...         self.observation_spec = CompositeSpec(observation=UnboundedContinuousTensorSpec(3))
...         self.reward_spec = UnboundedContinuousTensorSpec(1)
...         self.action_spec = UnboundedContinuousTensorSpec(1)
...
...     def _reset(self, tensordict: TensorDictBase, **kwargs) -> TensorDictBase:
...         tensordict_batch_size = tensordict.batch_size if tensordict is not None else torch.Size([])
...         result = self.observation_spec.rand(tensordict_batch_size)
...         result.update(self.full_done_spec.zero(tensordict_batch_size))
...         return result
...
...     def _step(
...         self,
...         tensordict: TensorDictBase,
...     ) -> TensorDictBase:
...         result = self.observation_spec.rand(tensordict.batch_size)
...         result.update(self.full_done_spec.zero(tensordict.batch_size))
...         result.update(self.full_reward_spec.zero(tensordict.batch_size))
...         return result
...
...     def _set_seed(self, seed: Optional[int]):
...         pass
...
>>> env = TransformedEnv(MyEnv(), BatchSizeTransform([5]))
>>> assert env.batch_size == torch.Size([5])
>>> assert env.rollout(10).shape == torch.Size([5, 10])

reset_func 可以创建具有所需批大小的 tensordict，从而允许进行细粒度的重置调用

>>> def reset_func(tensordict, tensordict_reset, env):
...     result = env.observation_spec.rand()
...     result.update(env.full_done_spec.zero())
...     assert result.batch_size != torch.Size([])
...     return result
>>> env = TransformedEnv(MyEnv(), BatchSizeTransform([5], reset_func=reset_func, env_kwarg=True))
>>> print(env.rollout(2))
TensorDict(
    fields={
        action: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.float32, is_shared=False),
        done: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.bool, is_shared=False),
        next: TensorDict(
            fields={
                done: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.bool, is_shared=False),
                observation: Tensor(shape=torch.Size([5, 2, 3]), device=cpu, dtype=torch.float32, is_shared=False),
                reward: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.float32, is_shared=False),
                terminated: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.bool, is_shared=False)},
            batch_size=torch.Size([5, 2]),
            device=None,
            is_shared=False),
        observation: Tensor(shape=torch.Size([5, 2, 3]), device=cpu, dtype=torch.float32, is_shared=False),
        terminated: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.bool, is_shared=False)},
    batch_size=torch.Size([5, 2]),
    device=None,
    is_shared=False)

此转换可用于在数据收集器中部署非批次锁定环境

>>> from torchrl.collectors import SyncDataCollector
>>> collector = SyncDataCollector(env, lambda td: env.rand_action(td), frames_per_batch=10, total_frames=-1)
>>> for data in collector:
...     print(data)
...     break
TensorDict(
    fields={
        action: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.float32, is_shared=False),
        collector: TensorDict(
            fields={
                traj_ids: Tensor(shape=torch.Size([5, 2]), device=cpu, dtype=torch.int64, is_shared=False)},
            batch_size=torch.Size([5, 2]),
            device=None,
            is_shared=False),
        done: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.bool, is_shared=False),
        next: TensorDict(
            fields={
                done: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.bool, is_shared=False),
                observation: Tensor(shape=torch.Size([5, 2, 3]), device=cpu, dtype=torch.float32, is_shared=False),
                reward: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.float32, is_shared=False),
                terminated: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.bool, is_shared=False)},
            batch_size=torch.Size([5, 2]),
            device=None,
            is_shared=False),
        observation: Tensor(shape=torch.Size([5, 2, 3]), device=cpu, dtype=torch.float32, is_shared=False),
        terminated: Tensor(shape=torch.Size([5, 2, 1]), device=cpu, dtype=torch.bool, is_shared=False)},
    batch_size=torch.Size([5, 2]),
    device=None,
    is_shared=False)
>>> collector.shutdown()

forward(tensordict: TensorDictBase) → TensorDictBase¶: 读取输入 tensordict，并针对选定的键应用转换。

transform_env_batch_size(batch_size: Size)[源代码]¶: 转换父环境的批大小。

transform_input_spec(input_spec: CompositeSpec) → CompositeSpec[源代码]¶

转换输入规范，使生成的规范匹配转换映射。

参数：: input_spec (TensorSpec) – 转换前的规范
返回值：: 转换后的预期规范

transform_output_spec(output_spec: CompositeSpec) → CompositeSpec[source]¶

转换输出规范，使结果规范与转换映射匹配。

此方法通常应保持不变。更改应使用 transform_observation_spec()、transform_reward_spec() 和 transformfull_done_spec() 实现。 :param output_spec: 变换前的规范 :type output_spec: TensorSpec

返回值：: 转换后的预期规范

BatchSizeTransform¶

文档

教程

资源