apache/beam · error · TypeError

batch type must be torch.Tensor or…

Error message

batch type must be torch.Tensor or beam.typehints.pytorch_type_compatibility.PytorchTensor[..]

What it means

PytorchBatchConverter.from_typehints builds a converter between individual PyTorch model elements and tensor batches. batch_type must be torch.Tensor or a PytorchTensor[...] Beam constraint; anything else raises this TypeError. This ensures batching operates only on genuine torch tensor types with a known dtype.

Solutions

1. Pass the actual torch.Tensor class object (import torch) or a PytorchTensor[dtype, shape] constraint as batch_type.
2. If batch_type comes from config as a string, resolve it to torch.Tensor before calling.
3. Also verify element_type is a PytorchTypeConstraint (or wrap it with PytorchTensor[element_type, ()]) and that element and batch dtypes match to avoid the follow-up dtype TypeError.
4. Catch TypeError at converter construction to fail fast before launching the pipeline.

Example fix

// before
converter = PytorchBatchConverter.from_typehints(element_type=model_output, batch_type='torch.Tensor')
// after
import torch
from apache_beam.typehints import pytorch_type_compatibility as ptc
converter = PytorchBatchConverter.from_typehints(
    element_type=model_output, batch_type=torch.Tensor)

import torch
from apache_beam.typehints.pytorch_type_compatibility import PytorchTensor
def valid_torch_batch_type(batch_type) -> bool:
    return batch_type == torch.Tensor or isinstance(batch_type, type(torch.Tensor[object, ()]) if False else object)

def is_torch_batch_type(batch_type) -> bool:
    import torch
    from apache_beam.typehints.pytorch_type_compatibility import PytorchTensor, PytorchTypeHint
    return batch_type == torch.Tensor or isinstance(batch_type, PytorchTypeHint.PytorchTypeConstraint)

try:
    converter = PytorchBatchConverter.from_typehints(element_type=et, batch_type=bt)
except TypeError as e:
    raise ValueError(f'Unsupported torch batch type {bt!r}; use torch.Tensor or PytorchTensor[..]') from e

When it happens

Trigger: Calling PytorchBatchConverter.from_typehints (via create_pytorch_batch_converter or BatchConverter.from_typehints) with batch_type set to something other than torch.Tensor or a PytorchTypeConstraint, e.g. numpy.ndarray, a string 'torch.Tensor', or a list type.

Common situations: Configuring batching for PyTorch inference pipelines where the batch type was configured from YAML/config as a string; mixing numpy and torch inference code; forgetting to import torch and passing a placeholder class.

Understand the failure class

Background: "Must be a positive integer", "Invalid value", "Unsupported": the invalid-argument-value error family, when a library rejects the value you pass — this error's family across 35 libraries.

Related errors

• A has been supplied to the model handler, but the required…
• batch type must be pa.Table or pa.Array
• batch type must be pd.Series or pd.DataFrame
• element_size_fn must be callable
• Element type must be compatible with Beam Schemas…

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/2a51368df677852a. Report an issue: GitHub.

Appendix: source

Thrown at sdks/python/apache_beam/typehints/pytorch_type_compatibility.py:49

      element_type,
      dtype,
      element_shape=(),
      partition_dimension=0):
    super().__init__(batch_type, element_type)
    self.dtype = dtype
    self.element_shape = element_shape
    self.partition_dimension = partition_dimension

  @staticmethod
  @BatchConverter.register(name="pytorch")
  def from_typehints(element_type,
                     batch_type) -> Optional['PytorchBatchConverter']:
    if not isinstance(element_type, PytorchTypeHint.PytorchTypeConstraint):
      element_type = PytorchTensor[element_type, ()]

    if not isinstance(batch_type, PytorchTypeHint.PytorchTypeConstraint):
      if not batch_type == torch.Tensor:
        raise TypeError(
            "batch type must be torch.Tensor or "
            "beam.typehints.pytorch_type_compatibility.PytorchTensor[..]")
      batch_type = PytorchTensor[element_type.dtype, (N, )]

    if not batch_type.dtype == element_type.dtype:
      raise TypeError(
          "batch type and element type must have equivalent dtypes "
          f"(batch={batch_type.dtype}, element={element_type.dtype})")
    computed_element_shape = list(batch_type.shape)
    partition_dimension = computed_element_shape.index(N)
    computed_element_shape.pop(partition_dimension)
    if not tuple(computed_element_shape) == element_type.shape:
      raise TypeError(
          "Could not align batch type's batch dimension with element type. "
          f"(batch type dimensions: {batch_type.shape}, element type "
          f"dimenstions: {element_type.shape}")

    return PytorchBatchConverter(

View on GitHub (pinned to 12126d8942)