Closed wsy-yjys closed 4 months ago
Your work is excellent, but there are two questions about the paper:
Question 1. Are all the comparison algorithms in Table 1 only half-trained epochs, such as 150 epochs for fair comparison?
Question 2. Are all the comparison algorithms in Table 2 only train 150 epochs?
As we stated in the experimental settings, we train our models as well as the reported DETR and RT-DETR for 150 epochs. APs of other detectors are directly cited from the original papers or github repos.
Your work is excellent, but there are two questions about the paper:
Question 1. Are all the comparison algorithms in Table 1 only half-trained epochs, such as 150 epochs for fair comparison?
Question 2. Are all the comparison algorithms in Table 2 only train 150 epochs?