how to understand the loss function corner case for 0 and 255

[weixsong]

Hi,

In the loss function, the code give a special value to value 0 and 255:

    # log probability for edge case of 0 (before scaling)
    log_cdf_plus = plus_in - tf.nn.softplus(plus_in)
    # log probability for edge case of 255 (before scaling)
    log_one_minus_cdf_min = -tf.nn.softplus(min_in)

I could not understand these code, how could me understand it?

[npuichigo]

I think you should know the relationship between sigmoid and softplus.

[weixsong]

@npuichigo , thanks very much. I got it.

[chediBechikh]

Hi all,
the explanation is not clear here and what relation what exactly used for this line

log probability for edge case of 0 (before scaling)

log_cdf_plus = plus_in - tf.nn.softplus(plus_in)