Something wrong with get_pixel_value?

[yasser-taha]

@kevinzakka @robotrory
In get_pixel_value, the description states that x and y are flattened tensors of shape (BxHxW,). However, in the first lines of the function, it is assumed that the input tensors are actually of the size (B,H,W,C) as shown here:

    shape = tf.shape(x)
    batch_size = shape[0]
    height = shape[1]
    width = shape[2]

This creates an error as shape[1] or higher is out of bound for a flattened tensor.

What am i missing?

[itxiud2015]

@YasserTahaSW Did you solved it?