Any suggestion about model ensemble for mask rcnn? #8728
-
|
I have trained several mask rcnn models with different backbones, then I want to ensemble these models for a better performance, but I am totally confused about how to do this for instance segmentation or object detection, can you give me some suggestion or reference for this? Thank you anyway. |
Beta Was this translation helpful? Give feedback.
Replies: 5 comments
-
|
I have the same problem. |
Beta Was this translation helpful? Give feedback.
-
|
I have the same problem. did you slove it? |
Beta Was this translation helpful? Give feedback.
-
|
I did guess no one did? |
Beta Was this translation helpful? Give feedback.
-
|
Hey. Mosty simple way is to combine all outputs and find a weighted average of it. |
Beta Was this translation helpful? Give feedback.
-
|
For boxes ensemble, you can simply nms the multi-model box outputs. See WBF for references. |
Beta Was this translation helpful? Give feedback.
For boxes ensemble, you can simply nms the multi-model box outputs.
See WBF for references.
https://github.com/ZFTurbo/Weighted-Boxes-Fusion