What is tensor Rank?

Question

I came through tensor slices and tensor rank. I read a lot of material related to this but could not satisfy myself.

Can anyone explain thoroughly with the help of examples, specifically in the view of TensorFlow?

score 1 · Accepted Answer · answered Mar 27 '21 at 13:21

1

tf.rank is like array.ndim in NumPy. It's the number of dimensions. E.g., a picture has rank 3:

640, 640, 3

A batched picture dataset has rank 4:

10000, 640, 640, 3

Tabular data has rank 2:

150, 4

It's that simple. Let me know if I should add anything.

answered Mar 27 '21 at 13:21

Nicolas Gervais

Amazing! Thanks alot – Ahmad Mar 27 '21 at 13:24
can you explain `dataset.shuffle(1000).repeat().batch(self.train_batch_size)` – Ahmad Mar 27 '21 at 14:28
https://stackoverflow.com/questions/64372390/im-trying-to-understand-what-buffer-size-in-dcgan-does/64372547#64372547 – Nicolas Gervais Mar 27 '21 at 15:22

1 Answers1