tensorflow/models · error · ValueError

The number of filters must be evenly divisible by the number

Error message

The number of filters must be evenly divisible by the number of groups. Received: groups={}, filters={}

What it means

Error "The number of filters must be evenly divisible by the number of groups. Received: groups={}, filters={}" thrown in tensorflow/models.

Source

Thrown at official/vision/modeling/layers/nn_layers.py:884

      pad_total.append(tf.maximum(kernel_size_effective[i] - overlap, 0))
    pad_beg = [pad_total[i] // 2 for i in range(self.rank)]
    pad_end = [pad_total[i] - pad_beg[i] for i in range(self.rank)]
    padding = [[pad_beg[i], pad_end[i]] for i in range(self.rank)]
    padding = [[0, 0]] + padding + [[0, 0]]

    if use_buffered_input:
      padding[time_axis] = [0, 0]
    else:
      padding[time_axis] = [padding[time_axis][0] + padding[time_axis][1], 0]
    return padding

  def _causal_validate_init(self):
    """Validates the Conv layer initial configuration."""
    # Overriding this method is meant to circumvent unnecessary errors when
    # using causal padding.
    if (self.filters is not None
        and self.filters % self.groups != 0):
      raise ValueError(
          'The number of filters must be evenly divisible by the number of '
          'groups. Received: groups={}, filters={}'.format(
              self.groups, self.filters))

    if not all(self.kernel_size):
      raise ValueError('The argument `kernel_size` cannot contain 0(s). '
                       'Received: %s' % (self.kernel_size,))

  def _buffered_spatial_output_shape(self, spatial_output_shape: List[int]):
    """Computes the spatial output shape from the input shape."""
    # When buffer padding, use 'valid' padding across time. The output shape
    # across time should be the input shape minus any padding, assuming
    # the stride across time is 1.
    if self._use_buffered_input and spatial_output_shape[0] is not None:
      padding = self._compute_buffered_causal_padding(
          tf.zeros([1] + spatial_output_shape + [1]), use_buffered_input=False)
      spatial_output_shape[0] -= sum(padding[1])
    return spatial_output_shape

View on GitHub (pinned to e006f5f0d5)

Solutions

1. Choose filters that is an integer multiple of groups.
2. Set groups to 1 or adjust filters so filters % groups == 0.

When it happens

Trigger: Thrown at official/vision/modeling/layers/nn_layers.py:884 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/b0ff5e5df4ec3c24. Report an issue: GitHub.