How to reproduce the results using GLIP

[LgQu]

Hi,

Thanks for the great work. I wonder how to reproduce the GLIP results in Table 2 of the paper. I ran the script GLIP_eval/eval.py and got the detection results saved in "glip_results.json", but had no idea how to calculate the GLIP results. Could you please give some advice to calculate the GLIP-related metrics, e.g., Zero/One obj. and Two obj.? Many thanks!

[weixi-feng]

Hi,

Thanks for the question. Since the prompts follow the template "ObjectA and ObjectB", you just need to compute the ratio of images where both ObjectA & ObjectB are detected by GLIP to get "Two obj." and the rest are "Zero/One obj.".

[LgQu]

Thanks for the clarification. Did you use the default 0.5 as the threshold?