About the issue of reproduction

[huamo555]

Hello Author:

Thank you for your outstanding work, I am now trying to reproduce your work, I trained 380 epochs using realsense camera and the three AP metrics are 65.72/62.26/31.14, which are far from the metrics in your paper. May I ask how many epochs you trained at that time?

[huamo555]

你好作者：

感谢您的出色工作，我现在正在尝试重现您的工作，我使用实感相机训练了 380 个 epoch，三个 AP 指标是 65.72/62.26/31.14，这与您论文中的指标相去甚远。请问您当时训练了多少个纪元？

I checked the tensorboard of the training process and it seems that the model has converged, I hope to get your reply in order to help me solve this problem, thank you very much.