DAG NCCL channel error when binding with a node of the same actor

channel_error.py

import ray
import torch
from ray.dag.input_node import InputNode
from ray.dag.output_node import MultiOutputNode
from ray.experimental.channel.torch_tensor_type import TorchTensorType

@ray.remote(num_gpus=1)
class MyActor:
    def __init__(self):
        pass

    def entrypoint(self, inp):
        pass

    def aggregate(self, *args):
        return args[0]

    def forward(self, inp):
        return torch.randn(10, 10).cuda()



workers = [MyActor.remote() for _ in range(2)]

with InputNode() as input_node:
    entrypoint = workers[0].entrypoint.bind(input_node)
    activations = [worker.forward.bind(entrypoint) for worker in workers]

    for activation in activations:
        activation.with_type_hint(
            TorchTensorType(transport=TorchTensorType.NCCL)
        )

    dag = workers[0].aggregate.bind(*activations)
    dag = dag.experimental_compile()
dag.execute(1)