ValueError: too many values to unpack (expected 4)

[StefanoGarzoni]

hi, i'm trying to run the code of this video: https://www.youtube.com/watch?v=QV85eYOb7gk&t=55s
after double-checking several times, I can say I'm sure it's identical to the example shown in the video, but when I go to run it in my anaconda environment (after obviously installing supervision and ultralytics) it gives me the following error:
0: 480x640 1tv, 814.3ms
Speed: 2.0ms preprocess, 814.3ms inference, 2.0ms postprocess for image at shape (1, 3, 640, 640)
Traceback (most recent call last):
File "C:\Users\stefano\Desktop\AI Project\Progetti\Real Time Counting Object\OpenCv\venv\countingRealTime.py", line 81, in
main()
File "C:\Users\stefano\Desktop\AI Project\Progetti\Real Time Counting Object\OpenCv\venv\countingRealTime.py", line 60, in main
labels = [
^
File "C:\Users\stefano\Desktop\AI Project\Progetti\Real Time Counting Object\OpenCv\venv\countingRealTime.py", line 62, in
for _, confidence, class_id, _
^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: too many values to unpack (expected 4)

I can't understand why.
Can anyone kindly help me? Thanks so much in advance everyone!

[SkalskiP]

Hi 👋🏻! Downgrade supervision version to 0.3.0.

[abra-abe]

for me this solution works but as soon as an object is detected the frame closes and this error is thrown again, is there another solution ? thank you in advance

[SkalskiP]

Hi, @abra-abe 👋🏻!

I need more information. Could you provide me with some code snippets?

[abra-abe]

I used the following code:

import cv2
import argparse

from ultralytics import YOLO
import supervision as sv
import numpy as np


ZONE_POLYGON = np.array([
    [0, 0],
    [0.5, 0],
    [0.5, 1],
    [0, 1]
])


def parse_arguments() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description="YOLOv8 live")
    parser.add_argument(
        "--webcam-resolution", 
        default=[640, 360], 
        nargs=2, 
        type=int
    )
    args = parser.parse_args()
    return args


def main():
    args = parse_arguments()
    frame_width, frame_height = args.webcam_resolution

    cap = cv2.VideoCapture("/dev/video2")
    cap.set(cv2.CAP_PROP_FRAME_WIDTH, frame_width)
    cap.set(cv2.CAP_PROP_FRAME_HEIGHT, frame_height)

    model = YOLO("best.pt")

    box_annotator = sv.BoxAnnotator(
        thickness=2,
        text_thickness=2,
        text_scale=1
    )

    zone_polygon = (ZONE_POLYGON * np.array(args.webcam_resolution)).astype(int)
    zone = sv.PolygonZone(polygon=zone_polygon, frame_resolution_wh=tuple(args.webcam_resolution))
    zone_annotator = sv.PolygonZoneAnnotator(
        zone=zone, 
        color=sv.Color.red(),
        thickness=2,
        text_thickness=4,
        text_scale=2
    )

    while True:
        ret, frame = cap.read()

        result = model(frame, agnostic_nms=True)[0]
        detections = sv.Detections.from_yolov8(result)
        labels = [
            f"{model.model.names[class_id]} {confidence:0.2f}"
            for _, confidence, class_id, _
            in detections
        ]
        frame = box_annotator.annotate(
            scene=frame, 
            detections=detections, 
            labels=labels
        )

        zone.trigger(detections=detections)
        frame = zone_annotator.annotate(scene=frame)      
        
        cv2.imshow("yolov8", frame)

        if (cv2.waitKey(3) == 27):
            break


if __name__ == "__main__":
    main()

It gives me the following output:

0: 608x800 (no detections), 801.3ms
Speed: 23.4ms preprocess, 801.3ms inference, 8.9ms postprocess per image at shape (1, 3, 608, 800)

0: 608x800 (no detections), 664.8ms
Speed: 13.1ms preprocess, 664.8ms inference, 0.8ms postprocess per image at shape (1, 3, 608, 800)

0: 608x800 (no detections), 561.3ms
Speed: 8.1ms preprocess, 561.3ms inference, 0.7ms postprocess per image at shape (1, 3, 608, 800)

0: 608x800 (no detections), 519.2ms
Speed: 12.7ms preprocess, 519.2ms inference, 0.6ms postprocess per image at shape (1, 3, 608, 800)

0: 608x800 (no detections), 453.7ms
Speed: 5.9ms preprocess, 453.7ms inference, 0.9ms postprocess per image at shape (1, 3, 608, 800)

0: 608x800 (no detections), 452.3ms
Speed: 6.5ms preprocess, 452.3ms inference, 0.6ms postprocess per image at shape (1, 3, 608, 800)

0: 608x800 1 plastic, 452.9ms
Speed: 6.3ms preprocess, 452.9ms inference, 17.1ms postprocess per image at shape (1, 3, 608, 800)
Traceback (most recent call last):
  File "/home/turtleg/Desktop/UNI SEASON 8 Series Finale/Project/control panel/test.py", line 81, in <module>
    main()
  File "/home/turtleg/Desktop/UNI SEASON 8 Series Finale/Project/control panel/test.py", line 60, in main
    labels = [
             ^
  File "/home/turtleg/Desktop/UNI SEASON 8 Series Finale/Project/control panel/test.py", line 62, in <listcomp>
    for _, confidence, class_id, _
        ^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: too many values to unpack (expected 4)

Sorry for the late reply I was just working with it through the CLI which is just working perfectly so far.

[SkalskiP]

Cool. But what version of supervision do you use?

[StefanoGarzoni]

Thanks so much, now it works!

[KHRMNKY]

I'm getting the same error too.
my supervision version : 0.18.0

#my code :

import cv2
from ultralytics import YOLO
import os
import supervision as sv


__file__= "HUGGINGFACE _MODELS_AND_SPACES"
current_dırectory= os.path.dirname(os.path.abspath(__file__))
folder= os.path.join(current_dırectory,"Cattle_Detection_with_YOLOV8")
pt= os.path.join(folder, "best.pt")
py= os.path.join(folder, "Detection_Video.py")
rqrmt= os.path.join(folder, "requirements.txt")
example_video= os.path.join(folder, "cows-and-cows-and-cows.mp4")
output_video= os.path.join(folder, "output_video.mp4")


def fonk(video_path):
    model=YOLO(pt)
    generator = sv.get_video_frames_generator(video_path)
    target_video = "/home/kahraman/Masaüstü/video detection/HuggingFace_Models_and_Spaces/Cattle_Detection_with_YOLOV8/cows-and-cows-and-cows.mp4"
    video_info = sv.VideoInfo.from_video_path(target_video)
    with sv.VideoSink(target_path=output_video, video_info=video_info) as sink:
        for frame in generator:
            results = model(frame)[0]
            detections = sv.Detections.from_ultralytics(results)
            box_annotator = sv.BoxAnnotator()
            classes = model.model.names 

            labels = [
            f"{detections[class_id]} {detections.confidence:0.2f}"
            for _, _, confidence, class_id, _  in detections
            ]

            annotated_image = box_annotator.annotate(frame, detections, labels)
            sink.write_frame(frame=annotated_image)
                

fonk("/home/kahraman/Masaüstü/video detection/HuggingFace_Models_and_Spaces/Cattle_Detection_with_YOLOV8/cows-and-cows-and-cows.mp4")

#output:

kahraman@kahraman-Lenovo-IdeaPad-S145-15IWL:~/Masaüstü/video detection/HuggingFace_Models_and_Spaces$ /bin/python3 "/home/kahraman/Masaüstü/video detection/HuggingFace_Models_and_Spaces/Cattle_Detection_with_YOLOV8/pratic1.py"

0: 640x384 3 cattles, 55.7ms
Speed: 10.5ms preprocess, 55.7ms inference, 15.1ms postprocess per image at shape (1, 3, 640, 384)
Traceback (most recent call last):
  File "/home/kahraman/Masaüstü/video detection/HuggingFace_Models_and_Spaces/Cattle_Detection_with_YOLOV8/pratic1.py", line 39, in <module>
    fonk("/home/kahraman/Masaüstü/video detection/HuggingFace_Models_and_Spaces/Cattle_Detection_with_YOLOV8/cows-and-cows-and-cows.mp4")
  File "/home/kahraman/Masaüstü/video detection/HuggingFace_Models_and_Spaces/Cattle_Detection_with_YOLOV8/pratic1.py", line 29, in fonk
    labels = [
  File "/home/kahraman/Masaüstü/video detection/HuggingFace_Models_and_Spaces/Cattle_Detection_with_YOLOV8/pratic1.py", line 31, in <listcomp>
    for _, _, confidence, class_id, _  in detections
ValueError: too many values to unpack (expected 5)

[ghost]

Did you solve this?

[SkalskiP]

Hi, @CarolinaClaro 👋🏻 are you also facing this issue? Can you show the code you are running and check what version of supervision you have installed?

[Nasarae]

Hey I'm the ghost 🥲, I totally forgot to answer this. I believe downgrading the version at the time worked. However I also remember debugging exactly what "detections" had in it (you probably say in the tutorial anyways). Thanks!