ray-project
[RLlib] TorchDistributionWrapper Typing Information Should Be Changed
What happened + What you expected to happen
The code functions but the typing for TorchDistributionWrapper's constructor in torch_action_dist.py should be improved. The code in question is:
def __init__(self, inputs: List[TensorType], model: TorchModelV2):
if not isinstance(inputs, torch.Tensor):
The argument typing says you have to pass a list and then on the next line the code checks if it's something that's not a list. Python's isinstance will not check members of collections so isinstance([1,2,3], int) will return False.
So either this function can be expecting inputs to be a Tensor in which case you probably want something like:
def __init__(self, inputs: Union[TensorType, List[TensorType]], model: TorchModelV2):
Or it doesn't in which case the if not isinstance check should be removed as it's always True as a List is not a Tensor.
Versions / Dependencies
This code is currently on HEAD and I found it reading the 2.2 source.
Reproduction script
This is a code/documentation quality/clarity problem, not a functionality issue. You just read the code.
Issue Severity
Low: It annoys or frustrates me.
This P2 issue has seen no activity in the past 2 years. It will be closed in 2 weeks as part of ongoing cleanup efforts.
Please comment and remove the pending-cleanup label if you believe this issue should remain open.
Thanks for contributing to Ray!
