Skip to content
This repository was archived by the owner on Jan 29, 2021. It is now read-only.

Model Input and Output

Yannis edited this page Jun 15, 2017 · 6 revisions

Input: Numpy array

Channels: RGB

Dimensions: 1 x 120 x 160 x 3

  • corresponds to an image of 160x120 pixels, RGB
  • Keras requires a four-dimensional array as input, hence the extra dim in front

Value Range (for each channel, not yet implemented): [-1, 1]

  • -1 equals to 0 in the 0-255 scale
  • 1 equals to 255 in the 0-255 scale

Clone this wiki locally