DiscreteIQLLoss

class torchrl.objectives.DiscreteIQLLoss(*args, **kwargs)

TorchRL 实现离散 IQL 损失。

在 “Offline Reinforcement Learning with Implicit Q-Learning” 中提出 https://arxiv.org/abs/2110.06169

参数:

• actor_network (ProbabilisticActor) – 随机策略

• qvalue_network (TensorDictModule) – Q(s, a) 参数化模型。

• value_network (TensorDictModule, optional) – V(s) 参数化模型。

关键字参数:

• action_space (str or TensorSpec) – 动作空间。必须是 "one-hot"、"mult_one_hot"、"binary" 或 "categorical" 之一，或相应规格的实例（torchrl.data.OneHot、torchrl.data.MultiOneHot、torchrl.data.Binary 或 torchrl.data.Categorical）。

• num_qvalue_nets (integer, optional) – 使用的 Q 值网络的数量。默认为 2。

• loss_function (str, optional) – 要用于值函数损失的损失函数。默认为 “smooth_l1”。

• temperature (float, optional) – 逆温度（beta）。对于较小旳超参数值，该目标函数行为类似于行为克隆，而对于较大旳值，它试图恢复 Q 函数旳最大值。

• expectile (float, optional) – expectile \(\tau\)。对于需要动态规划（“stichting”）的 antmaze 任务，较大的 \(\tau\) 值至关重要。

• priority_key (str, optional) – [已弃用，请改用 .set_keys(priority_key=priority_key)] 写入优先级的 tensordict 键（用于优先回放缓冲区）。默认为 “td_error”。

• separate_losses (bool, optional) – 如果为 True，则策略和 critic 之间的共享参数将仅在策略损失上训练。默认为 False，即梯度同时为策略和 critic 损失传播到共享参数。

• reduction (str, optional) – 指定应用于输出的归约："none" | "mean" | "sum"。"none"：不应用归约，"mean"：输出的总和将除以输出中的元素数量，"sum"：将对输出求和。默认为 "mean"。

示例

>>> import torch
>>> from torch import nn
>>> from torchrl.data.tensor_specs import OneHot
>>> from torchrl.modules.distributions.discrete import OneHotCategorical
>>> from torchrl.modules.tensordict_module.actors import ProbabilisticActor
>>> from torchrl.modules.tensordict_module.common import SafeModule
>>> from torchrl.objectives.iql import DiscreteIQLLoss
>>> from tensordict import TensorDict
>>> n_act, n_obs = 4, 3
>>> spec = OneHot(n_act)
>>> module = SafeModule(nn.Linear(n_obs, n_act), in_keys=["observation"], out_keys=["logits"])
>>> actor = ProbabilisticActor(
...     module=module,
...     in_keys=["logits"],
...     out_keys=["action"],
...     spec=spec,
...     distribution_class=OneHotCategorical)
>>> qvalue = SafeModule(
...     nn.Linear(n_obs, n_act),
...     in_keys=["observation"],
...     out_keys=["state_action_value"],
... )
>>> value = SafeModule(
...     nn.Linear(n_obs, 1),
...     in_keys=["observation"],
...     out_keys=["state_value"],
... )
>>> loss = DiscreteIQLLoss(actor, qvalue, value)
>>> batch = [2, ]
>>> action = spec.rand(batch).long()
>>> data = TensorDict({
...         "observation": torch.randn(*batch, n_obs),
...         "action": action,
...         ("next", "done"): torch.zeros(*batch, 1, dtype=torch.bool),
...         ("next", "terminated"): torch.zeros(*batch, 1, dtype=torch.bool),
...         ("next", "reward"): torch.randn(*batch, 1),
...         ("next", "observation"): torch.randn(*batch, n_obs),
...     }, batch)
>>> loss(data)
TensorDict(
    fields={
        entropy: Tensor(shape=torch.Size([]), device=cpu, dtype=torch.float32, is_shared=False),
        loss_actor: Tensor(shape=torch.Size([]), device=cpu, dtype=torch.float32, is_shared=False),
        loss_qvalue: Tensor(shape=torch.Size([]), device=cpu, dtype=torch.float32, is_shared=False),
        loss_value: Tensor(shape=torch.Size([]), device=cpu, dtype=torch.float32, is_shared=False)},
    batch_size=torch.Size([]),
    device=None,
    is_shared=False)

此类与非 tensordict 的模块兼容，无需调用任何 tensordict 相关原语即可使用。在这种情况下，预期的关键字参数是：["action", "next_reward", "next_done", "next_terminated"] + actor、value 和 qvalue 网络的 in_keys。返回值是一个按以下顺序排列的张量元组：["loss_actor", "loss_qvalue", "loss_value", "entropy"]。

示例

>>> import torch
>>> import torch
>>> from torch import nn
>>> from torchrl.data.tensor_specs import OneHot
>>> from torchrl.modules.distributions.discrete import OneHotCategorical
>>> from torchrl.modules.tensordict_module.actors import ProbabilisticActor
>>> from torchrl.modules.tensordict_module.common import SafeModule
>>> from torchrl.objectives.iql import DiscreteIQLLoss
>>> _ = torch.manual_seed(42)
>>> n_act, n_obs = 4, 3
>>> spec = OneHot(n_act)
>>> module = SafeModule(nn.Linear(n_obs, n_act), in_keys=["observation"], out_keys=["logits"])
>>> actor = ProbabilisticActor(
...     module=module,
...     in_keys=["logits"],
...     out_keys=["action"],
...     spec=spec,
...     distribution_class=OneHotCategorical)
>>> qvalue = SafeModule(
...     nn.Linear(n_obs, n_act),
...     in_keys=["observation"],
...     out_keys=["state_action_value"],
... )
>>> value = SafeModule(
...     nn.Linear(n_obs, 1),
...     in_keys=["observation"],
...     out_keys=["state_value"],
... )
>>> loss = DiscreteIQLLoss(actor, qvalue, value)
>>> batch = [2, ]
>>> action = spec.rand(batch).long()
>>> loss_actor, loss_qvalue, loss_value, entropy = loss(
...     observation=torch.randn(*batch, n_obs),
...     action=action,
...     next_done=torch.zeros(*batch, 1, dtype=torch.bool),
...     next_terminated=torch.zeros(*batch, 1, dtype=torch.bool),
...     next_observation=torch.zeros(*batch, n_obs),
...     next_reward=torch.randn(*batch, 1))
>>> loss_actor.backward()

输出键也可以使用 DiscreteIQLLoss.select_out_keys() 方法进行过滤。

示例

>>> _ = loss.select_out_keys('loss_actor', 'loss_qvalue', 'loss_value')
>>> loss_actor, loss_qvalue, loss_value = loss(
...     observation=torch.randn(*batch, n_obs),
...     action=action,
...     next_done=torch.zeros(*batch, 1, dtype=torch.bool),
...     next_terminated=torch.zeros(*batch, 1, dtype=torch.bool),
...     next_observation=torch.zeros(*batch, n_obs),
...     next_reward=torch.randn(*batch, 1))
>>> loss_actor.backward()

default_keys

别名：_AcceptedKeys

forward(tensordict: TensorDictBase = None) → TensorDictBase

它旨在读取一个输入的 TensorDict 并返回另一个包含名为“loss*”的损失键的 tensordict。

将损失分解为其组成部分可以被训练器用于在训练过程中记录各种损失值。输出 tensordict 中存在的其他标量也将被记录。

参数:

tensordict – 一个输入的 tensordict，包含计算损失所需的值。

返回:

一个没有批处理维度的新 tensordict，其中包含各种损失标量，这些标量将被命名为“loss*”。重要的是，损失必须以这个名称返回，因为它们将在反向传播之前被训练器读取。