RandomPolicy¶
- torchrl.envs.utils.RandomPolicy(action_spec: TensorSpec, action_key: NestedKey = 'action')[source]¶
A random policy for data collectors.
This is a wrapper around the action_spec.rand method.
- Parameters:
action_spec – TensorSpec object describing the action specs
Examples
>>> from tensordict import TensorDict >>> from torchrl.data.tensor_specs import Bounded >>> action_spec = Bounded(-torch.ones(3), torch.ones(3)) >>> actor = RandomPolicy(action_spec=action_spec) >>> td = actor(TensorDict({}, batch_size=[])) # selects a random action in the cube [-1; 1]