FoundationVision / GLEE

[CVPR2024 Highlight]GLEE: General Object Foundation Model for Images and Videos at Scale
https://glee-vision.github.io/
MIT License
1.02k stars 82 forks source link

COCO performance after finetuning #14

Closed HaisongDing closed 5 months ago

HaisongDing commented 5 months ago

Have you tried to finetune the stage 2 model on COCO dataset and see the limit of the COCO performance?

wjf5203 commented 5 months ago

Yes, we have observed that continuing to finetune GLEE-Lite-Joint on COCO does not improve performance. We believe this highlights the advantage of GLEE—enhancing the model's detection abilities by training with a larger amount of data. However, the performance of GLEE-Pro on COCO has not yet met expectations, and we are currently exploring the reasons for this.