conflicting dependencies

[Faezehsgh]

it seems that different version of requirements of the project conflict with each other. for example the gym and the ray library how did you solve these conflicts?

[Reytuag]

Hey sorry for the late response, I was on vacation and then forgot about it. Hmm I think we forced the install of the version anyway and it was working if I remember correctly. Tell me if this does not work for you I can try to reinstall it and run it.

Also while looking for any info on this ( Because the code was started by an intern in our team) , He also mentions a bugfix in rllib he added (for the version of the config).

In ray/rllib/utils/torch_utils.py replace convert_to_torch_tensor by this definition

def convert_to_torch_tensor(x: TensorStructType, device: Optional[str] = None):
    """Converts any struct to torch.Tensors.

    x (any): Any (possibly nested) struct, the values in which will be
        converted and returned as a new struct with all leaves converted
        to torch tensors.

    Returns:
        Any: A new struct with the same structure as `stats`, but with all
            values converted to torch Tensor types.
    """

    def mapping(item):
        # ELIAS ADD THIS BUGFIX
        if item is None:
            # returns None with dtype=np.obj
            return np.asarray(item)

        # Already torch tensor -> make sure it's on right device.
        if torch.is_tensor(item):
            return item if device is None else item.to(device)
        # Special handling of "Repeated" values.
        elif isinstance(item, RepeatedValues):
            return RepeatedValues(
                tree.map_structure(mapping, item.values), item.lengths, item.max_len
            )
        # Numpy arrays.
        if isinstance(item, np.ndarray):
            # Object type (e.g. info dicts in train batch): leave as-is.
            if item.dtype == object:
                return item
            # Non-writable numpy-arrays will cause PyTorch warning.
            elif item.flags.writeable is False:
                with warnings.catch_warnings():
                    warnings.simplefilter("ignore")
                    tensor = torch.from_numpy(item)
            # Already numpy: Wrap as torch tensor.
            else:
                tensor = torch.from_numpy(item)
        # Everything else: Convert to numpy, then wrap as torch tensor.
        else:
            tensor = torch.from_numpy(np.asarray(item))
        # Floatify all float64 tensors.
        if tensor.dtype == torch.double:
            tensor = tensor.float()
        return tensor if device is None else tensor.to(device)

    return tree.map_structure(mapping, x)

Don't hesitate if you have further issues or question, I will add those in the readme