Code for "hyp_mlr"

[JinhwiPark]

Hi, thanks for the interesting paper.

I think the following part is the code for hyperbolic logit (paper equation(7)), and it seems that there is no $\lambda_{p_{y}}^{c}$ in the code:

HyperbolicImageSegmentation/hesp/util/layers.py

Line 115 in 622a173

return 2.0 / tf.sqrt(c) * A_norm * tf.asinh(sineterm)

Could you check this?

[seanzhuh]

Hi, I've encountered the same issue, and another problem I found out was that the input to tf.asinh does not divide w_y also

[jhindel]

I agree with @seanzhuh and @JinhwiPark. lambda is always set to 2 and w_y seems to be missing in the denominator. Is there any reason for it?