How to show Average Precision of Each Class

[Jethro-241]

In our object detection model, there's 2 classes: Paper and Plastic. How can I show the average precision of paper and plastic separately and their mAP?

what exception? All I get is mAP, P, R, etc. like this below:

What I want to get is this:

I want to get the average precision

Can you demonstrate it to me? Here's the colab link: https://colab.research.google.com/drive/10KL3fJbImLXmMAw8aSLOLlngQsQ6MG5P?usp=sharing

[WongKinYiu]

for your reference. https://medium.datadriveninvestor.com/how-to-create-custom-coco-data-set-for-object-detection-96ec91958f36

i do not find the test script in your colab.

[Jethro-241]

do you have an email address so we can send our test.py?

[WongKinYiu]

just prepare coco style gt and make sure it is put in the corresponding location as what in test.py, you will get the expected results.

[Jethro-241]

what line in your test.py will you see the corresponding location? so we can check it out in your code.

[WongKinYiu]

please install pycocotools and put coco style gt file as https://github.com/WongKinYiu/ScaledYOLOv4/blob/yolov4-large/test.py#L228 and set opt.save_json = True https://github.com/WongKinYiu/ScaledYOLOv4/blob/yolov4-large/test.py#L264

[Jethro-241]

thanks we will try it right now

[qutyyds]

How did you solve this? I have the same problem, too. i want to see per class but only can see "all". what should i do?

[WongKinYiu]

add following code after https://github.com/WongKinYiu/ScaledYOLOv4/blob/yolov4-large/test.py#L235

            for cid in range(80):
                print(names[cid])
                cocoEval.params.catIds = [coco91class[cid]]
                cocoEval.evaluate()
                cocoEval.accumulate()
                cocoEval.summarize()