tensorflow/models · error · ValueError

The token budget, global batch size, is too small to yield 0

Error message

The token budget, global batch size, is too small to yield 0 bucket window: %s

What it means

Error "The token budget, global batch size, is too small to yield 0 bucket window: %s" thrown in tensorflow/models.

Source

Thrown at official/nlp/data/wmt_dataloader.py:119

    batch_size: Max number of tokens per batch of examples.
    max_length: Max number of tokens in an example input or target sequence.

  Returns:
    Dataset of batched examples with similar lengths.
  """
  # Get min and max boundary lists for each example. These are used to calculate
  # the `bucket_id`, which is the index at which:
  # buckets_min[bucket_id] <= len(example) < buckets_max[bucket_id]
  # Note that using both min and max lists improves the performance.
  buckets_min, buckets_max = _create_min_max_boundaries(max_length)

  # Create list of batch sizes for each bucket_id, so that
  # bucket_batch_size[bucket_id] * buckets_max[bucket_id] <= batch_size
  bucket_batch_sizes = [int(batch_size) // x for x in buckets_max]

  # Validates bucket batch sizes.
  if any([batch_size <= 0 for batch_size in bucket_batch_sizes]):
    raise ValueError(
        'The token budget, global batch size, is too small to yield 0 bucket '
        'window: %s' % str(bucket_batch_sizes))

  # bucket_id will be a tensor, so convert this list to a tensor as well.
  bucket_batch_sizes = tf.constant(bucket_batch_sizes, dtype=tf.int64)

  def example_to_bucket_id(example):
    """Return int64 bucket id for this example, calculated based on length."""
    example_input = example['inputs']
    example_target = example['targets']
    seq_length = _get_example_length((example_input, example_target))

    conditions_c = tf.logical_and(
        tf.less_equal(buckets_min, seq_length), tf.less(seq_length,
                                                        buckets_max))
    bucket_id = tf.reduce_min(tf.where(conditions_c))
    return bucket_id

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/data/wmt_dataloader.py:119 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of tensorflow/models@e006f5f0d5 (2026-08-24). Data as JSON: /api/errors/6e110d4a06225df6. Report an issue: GitHub.