CODEWITHSUNDEEP
classinferenceyolov8
Dervan
Dervan

Reputation: 1

YOLOv8 Inference

I would like yolov8 to display the sum of each of the classes in an image on the CLI.

In yolov7 for example, when I run inference on a custom data set it displays something like this:

12 capacitor-sam2s, 5 capacitor-mur1s, 5 capacitor-mur2s, 1 rfid, 1 ntc, 2 resistor-packs, Done. (1513.7ms) Inference, (55.4ms) NMS

How do I do this with yolov8?

It doesn't display the results on the command line.

Upvotes: 0

Views: 233

Answers (1)

Natrium2
Natrium2

Reputation: 85

I would recommend you to achieve this in python and not via bash.

You have to route the corresponding labels to a dictionary.

You can do it like this:

model = YOLO(MODEL_NAME)
results = model.predict(imgsz=1024, source=filepath, save=False, conf=min_conf_threshold) # source already setup
dictionary = {}   
for r in results:
    for c in r.boxes.cls:
        labelnames = model.names[int(c)]
        dictionary[labelnames ] = dictionary.get(labelnames, 0) + 1

and afterwards you can simply sum it up over all classes:

amount = sum(dictionary.values())
dictionary["Cumulated"]= amount
for i in dictionary:
    print (i + str(":"), dictionary[i])

Upvotes: 1

Related Questions