WZMIAOMIAO/deep-learning-for-image-processing · error · ValueError

sampler should be an instance of torch.utils.data.Sampler, b

Error message

sampler should be an instance of torch.utils.data.Sampler, but got sampler={}

What it means

BatchSampler-style __init__ validates that the sampler argument is an instance of torch.utils.data.Sampler and raises ValueError otherwise, echoing the received object. This GroupedBatchSampler groups indices by aspect ratio, so it depends on the Sampler protocol (iter/len) to produce batches.

Source

Thrown at pytorch_keypoint/HRNet/train_utils/group_by_aspect_ratio.py:39

class GroupedBatchSampler(BatchSampler):
    """
    Wraps another sampler to yield a mini-batch of indices.
    It enforces that the batch only contain elements from the same group.
    It also tries to provide mini-batches which follows an ordering which is
    as close as possible to the ordering from the original sampler.
    Arguments:
        sampler (Sampler): Base sampler.
        group_ids (list[int]): If the sampler produces indices in range [0, N),
            `group_ids` must be a list of `N` ints which contains the group id of each sample.
            The group ids must be a continuous set of integers starting from
            0, i.e. they must be in the range [0, num_groups).
        batch_size (int): Size of mini-batch.
    """
    def __init__(self, sampler, group_ids, batch_size):
        if not isinstance(sampler, Sampler):
            raise ValueError(
                "sampler should be an instance of "
                "torch.utils.data.Sampler, but got sampler={}".format(sampler)
            )
        self.sampler = sampler
        self.group_ids = group_ids
        self.batch_size = batch_size

    def __iter__(self):
        buffer_per_group = defaultdict(list)
        samples_per_group = defaultdict(list)

        num_batches = 0
        for idx in self.sampler:
            group_id = self.group_ids[idx]
            buffer_per_group[group_id].append(idx)
            samples_per_group[group_id].append(idx)
            if len(buffer_per_group[group_id]) == self.batch_size:
                yield buffer_per_group[group_id]

View on GitHub (pinned to 1ec3fe6f37)

Solutions

1. Wrap plain indices in torch.utils.data.sampler.SubsetRandomSampler or RandomSampler before passing.
2. If you have aspect-ratio grouped data, build the sampler with create_aspect_ratio_groups() from the same module, which returns a proper Sampler.
3. Ensure you pass the sampler positionally in the right order (sampler, group_ids, batch_size).
4. If using torchvision's aspect ratio grouping, prefer torchvision.utils.dataset grouping utilities matching your torch version.

Example fix

# before
sampler = list(range(len(dataset)))
batch_sampler = GroupedBatchSampler(sampler, group_ids, batch_size)  # ValueError
# after
from torch.utils.data.sampler import SubsetRandomSampler
sampler = SubsetRandomSampler(list(range(len(dataset))))
batch_sampler = GroupedBatchSampler(sampler, group_ids, batch_size)

from torch.utils.data import Sampler
assert isinstance(sampler, Sampler), f"got {type(sampler).__name__}; wrap indices in a Sampler"

from torch.utils.data import Sampler
def is_valid_sampler(obj) -> bool:
    return isinstance(obj, Sampler)

try:
    batch_sampler = GroupedBatchSampler(sampler, group_ids, batch_size)
except ValueError as e:
    logging.error("Bad sampler: %s — wrapping in SubsetRandomSampler", e)
    batch_sampler = GroupedBatchSampler(SubsetRandomSampler(sampler), group_ids, batch_size)

When it happens

Trigger: Passing a plain list, range, generator, or a non-Sampler iterable as the first argument to the grouped batch sampler constructor, e.g. GroupedBatchSampler(dataset_indices, group_ids, batch_size) instead of a Sampler wrapping them.

Common situations: Upgrading PyTorch: newer torchvision versions made aspect-ratio grouping samplers subclass differently and older call sites pass indices directly; refactoring code that previously iterated raw index lists; confusing batch_sampler with sampler arguments in DataLoader.

Related errors

• Expected target boxes to be of type Tensor, got {:}.
• sampler should be an instance of torch.utils.data.Sampler, b
• sampler should be an instance of torch.utils.data.Sampler, b
• sampler should be an instance of torch.utils.data.Sampler, b
• return_layers are not present in model

AI-assisted analysis of WZMIAOMIAO/deep-learning-for-image-processing@1ec3fe6f37 (2026-08-30). Data as JSON: /api/errors/7fb45c80d7aca05b. Report an issue: GitHub.