search

Wayfear / Coacher

Other
22 stars 2 forks source link

About your evaluation metric zR@K #1

Closed yuxha closed 3 years ago

yuxha commented 3 years ago

Thanks for your work!

I am confused when I compare your experimental results with the results of your base code (SGG Benchmark Project). https://github.com/KaihuaTang/Scene-Graph-Benchmark.pytorch/blob/master/METRICS.md#explanation-of-our-metrics

I notice that the zR@Ks in your paper are much more than the zR@Ks in SGG Benchmark Project (About 10 times)

For example, the zR@20 of NeuralMotifs reported in SGG Benchmark Project is 1%, but 13% in your paper.

Can you explain why? Thank you very much.

ZHUXUHAN commented 2 years ago

Thanks for your work!

I am confused when I compare your experimental results with the results of your base code (SGG Benchmark Project). https://github.com/KaihuaTang/Scene-Graph-Benchmark.pytorch/blob/master/METRICS.md#explanation-of-our-metrics

I notice that the zR@Ks in your paper are much more than the zR@Ks in SGG Benchmark Project (About 10 times)

For example, the zR@20 of NeuralMotifs reported in SGG Benchmark Project is 1%, but 13% in your paper.

Can you explain why? Thank you very much.

hi, do you know why?