search

THU-MIG / yolov10

YOLOv10: Real-Time End-to-End Object Detection
https://arxiv.org/abs/2405.14458
GNU Affero General Public License v3.0
9.43k stars 888 forks source link

how to do inference with one2many #349

Open xmczh003 opened 2 months ago

xmczh003 commented 2 months ago

I noticed that the paper reported that using o2m yields a 0.6% improvement over o2o. However, in my usage scenario, I am less concerned about inference speed and more focused on inference accuracy. How can I use o2m alone?

qianyue76 commented 2 months ago

You can fix or rewrite the postprocess function in YOLOv10DetectionPredictor