Function sparse_categorical_crossentropy

tensorflow/python/keras/backend.py:4377–4451

Categorical crossentropy with integer targets. Arguments: target: An integer tensor. output: A tensor resulting from a softmax (unless `from_logits` is True, in which case `output` is expected to be the logits). from_logits: Boolean, whether `output` is the

(target, output, from_logits=False, axis=-1)

@keras_export('keras.backend.sparse_categorical_crossentropy')
def sparse_categorical_crossentropy(target, output, from_logits=False, axis=-1):
  """Categorical crossentropy with integer targets.

  Arguments:
      target: An integer tensor.
      output: A tensor resulting from a softmax
          (unless `from_logits` is True, in which
          case `output` is expected to be the logits).
      from_logits: Boolean, whether `output` is the
          result of a softmax, or is a tensor of logits.
      axis: Int specifying the channels axis. `axis=-1` corresponds to data
          format `channels_last', and `axis=1` corresponds to data format
          `channels_first`.

  Returns:
      Output tensor.

  Raises:
      ValueError: if `axis` is neither -1 nor one of the axes of `output`.
  """
  if not from_logits:
    if (isinstance(output, (ops.EagerTensor, variables_module.Variable)) or
        output.op.type != 'Softmax'):
      epsilon_ = _constant_to_tensor(epsilon(), output.dtype.base_dtype)
      output = clip_ops.clip_by_value(output, epsilon_, 1 - epsilon_)
      output = math_ops.log(output)
    else:
      # When softmax activation function is used for output operation, we
      # use logits from the softmax function directly to compute loss in order
      # to prevent collapsing zero when training.
      # See b/117284466
      assert len(output.op.inputs) == 1
      output = output.op.inputs[0]

  if isinstance(output.shape, (tuple, list)):
    output_rank = len(output.shape)
  else:
    output_rank = output.shape.ndims
  if output_rank is not None:
    axis %= output_rank
    if axis != output_rank - 1:
      permutation = list(
          itertools.chain(range(axis), range(axis + 1, output_rank), [axis]))
      output = array_ops.transpose(output, perm=permutation)
  elif axis != -1:
    raise ValueError(
        'Cannot compute sparse categorical crossentropy with `axis={}` on an '
        'output tensor with unknown rank'.format(axis))

  target = cast(target, 'int64')

  # Try to adjust the shape so that rank of labels = 1 - rank of logits.
  output_shape = array_ops.shape_v2(output)
  target_rank = target.shape.ndims

  update_shape = (
      target_rank is not None and output_rank is not None and
      target_rank != output_rank - 1)