Should the color values on the classification report be stretched from 0 to 1?

[braymp]

Since the lower bound of the colormap is set to the lowest value (and not to 0), sometimes I'm alarmed that there's a lot of variation in the colormap even though the actual values may not actually be much different.

Setting it from 0 to 1 gives a good visual perspective of the real variation. I think this is already being done for the confusion matrix...?