search

ShechemKS / Yolo_Detectron2

Implementation of Yolo using Facebook's Detectron2 (https://github.com/facebookresearch/detectron2) Framework with Quantization support based on AQD: Towards Accurate Quantized Object Detection (https://github.com/aim-uofa/model-quantization)
Apache License 2.0
9 stars 5 forks source link

The result is not right #1

Closed 1145520074 closed 2 years ago

1145520074 commented 3 years ago

I use this code to train coco dataset, but I found the result is not right.The mAP is bigger than 1 and the inferrence result is not wright.Thank you very much

AP AP50 AP75 APs APm APl
46.037 62.991 52.806 25.050 54.041 60.809

image

ShechemKS commented 3 years ago

Could you include your training and testing configuration files? The table with the results seems to be fine, are those the results that you obtained upon training? I can't figure out what is wrong without more information on your training/testing settings.

Also, it seems as though you have not run NMS on the results used to produce this image. Run NMS and see if the results are correct.

ShechemKS commented 2 years ago

Closing due to inactivity