4

The current version of Tensorflow Argmax doesn't specify what's the difference between the "axis" and "dimension" arguments. Here is the only information given in the official manual:

tf.argmax(input, axis=None, name=None, dimension=None) {#argmax}

Returns the index with the largest value across axes of a tensor.

Args:

input: A Tensor. Must be one of the following types: float32, float64, int64, int32, uint8, uint16, int16, int8, complex64, complex128, qint8, quint8, qint32, half.

axis: A Tensor. Must be one of the following types: int32, int64. int32, 0 <= axis < rank(input). Describes which axis of the input Tensor to reduce across. For vectors, use axis = 0.

name: A name for the operation (optional).

Returns:

A Tensor of type int64.

Can someone clarify? Which one is actually the dimension being reduced?

Dr_Zaszuś
  • 546
  • 1
  • 7
  • 15
  • 2
    Since `dimension` is not documented it means it's not part of API (ie, probably an old keyword arg that was left over from refactoring/renaming, do not use) – Yaroslav Bulatov Feb 22 '17 at 17:50

1 Answers1

5

TensorFlow is transitioning to use axis rather than dimension which is going to be deprecated soon: https://www.tensorflow.org/install/migration

yuefengz
  • 3,338
  • 1
  • 17
  • 24