Why yolox-s in mmyolo can be 40.8%(mAP)? However, official mAP of yolox-s is 40.5%?What you have modified?
lixiangMindSpore opened this issue · comments
Prerequisite
- I have searched the existing and past issues but cannot get the expected help.
- I have read the FAQ documentation but cannot get the expected help.
- The bug has not been fixed in the latest version.
💬 Describe the reimplementation questions
Why yolox-s in mmyolo can be 40.8%(mAP)? However, official mAP of yolox-s is 40.5%?What you have modified?
Environment
nothing special
Expected results
No response
Additional information
No response
@lixiangMindSpore It mainly includes 2 aspects:
- Using multi_label, the performance can be about 0.2
- Algorithms fluctuate
@lixiangMindSpore It mainly includes 2 aspects:
- Using multi_label, the performance can be about 0.2
- Algorithms fluctuate
How to use multi_label?
@lixiangMindSpore
Please refer to this line in the config file:
@lixiangMindSpore Please refer to this line in the config file:
Official YOLOX don't possess it. How to modify it in Official YOLOX?
@lixiangMindSpore
In MMYOLO, it's the same setting.
https://github.com/open-mmlab/mmyolo/blob/main/configs/yolox/yolox_s_8xb8-300e_coco.py#L97
If you want to modify official YOLOX, you can go to the official YOLOX repo and open a new issue about it.