The tunable parameters of VPT+Bias for Semantic Segmentation

[liulingbo918]

Can you provide more details of different methods and their tunable parameters in Table 4 for semantic segmentation?

Except for header parameters： 1）the tunable parameter number of BIAS is 13.46-13.18=0.28M 2）the tunable parameter number of VPT is 13.43-13.18=0.25M 3) why is the tunable parameter number of VPT+BIAS 15.79-13.18=2.61M, rather than 0.28+0.25=0.53M？

It seems to me that BIAS was reimplemented based on the paper [5] (fine-tunes only the bias terms). However, was VPT+BIAS reimplemented based on the paper [8] (fine-tunes the bias terms and introduces some lightweight residual layers)?

[KMnP]

Hi @liulingbo918 Thanks for your question. The reason for this difference is that VPT-deep and VPT+Bias use different number of prompts as hyperparameter. VPT-deep uses p=10, whereas VPT+BIAS uses p=100. Since VPT+BIAS use 10 times of prompts than VPT_deep, the tunable parameters of VPT+BIAS = 0.28 + 0.25 x10

Let me know if you have more questions!

[liulingbo918]

@KMnP Thanks for your reply. A good work!