What value of m (ID_MARGIN) was used across the paper? & What output scaling to use?

[majedelhelou]

Hello

1. Comparing the results in Table 1 to those in Table 4, it seems that m=0 was used for Table 1. I am trying to reproduce your results, is the value of m always 0 for all experiments in the paper (ignoring Figure 3)? What is the difference between: "Ours (ID+attributes)" and "Ours" in Table 1?

2. In your inversion and anonymization scripts, when saving images you use the function tensor2image that has the adaptive option set to true by default. This uses adaptive scaling when scaling tensors to images and causes somewhat washed out images. Did you use adaptive scaling for the results you show in the paper / should I keep the adaptive scaling on when using your model or not?

Thank you.