torchrl.trainers.algorithms.configs.modules.ValueModelConfig

class torchrl.trainers.algorithms.configs.modules.ValueModelConfig(_partial_: bool = False, in_keys: Optional[Any] = None, out_keys: Optional[Any] = None, _target_: str = 'torchrl.trainers.algorithms.configs.modules._make_value_model', network: NetworkConfig = '???')

用于配置 Value 模型的类。

示例

>>> cfg = ValueModelConfig(network=MLPConfig(in_features=10, out_features=5, depth=2, num_cells=32))
>>> net = instantiate(cfg)
>>> y = net(torch.randn(1, 10))
>>> assert y.shape == (1, 5)

另请参阅

torchrl.modules.ValueOperator