problem about config of benchmark

hustvl / MIMDet

[ICCV 2023] You Only Look at One Partial Sequence

https://arxiv.org/abs/2204.02964

MIT License

330 stars 30 forks source link

problem about config of benchmark #5

Closed UcanSee closed 2 years ago

UcanSee commented 2 years ago

Thanks for sharing such a good work! Now I want to use your codes to reproduce performance of mask-rcnn-benchmarking-ep50, but it seems that the config is not consistent with the paper. In paper, bs=64, lr=8.0e-5, wd=0.1 but bs=16, lr=1.6e-4, wd=0.2 in your config which is not consistent with the train log you released. Is there something I understand wrongly? Could you please provide an exact config?

vealocia commented 2 years ago

Hi, @UcanSee ! Thanks for your interest in our work. We have already updated the configs of benchmarking models few hours ago, please check the newest config files.

Yuxin-CV commented 2 years ago

I believe the issue at hand was addressed, as such I'm closing this. Feel free to ask if you have further questions.