Need help. Boxes galore of False Positives

Support YOLO
1

Hey everyone. I am new to Yolov8 and all of this. I am doing my best to learn and have watch different videos how to get this to work properly. When I test a pretrained model, I input an image and its output has literally every box and label on the output.
image
It’s an image of me holding a bottle.
Here’s my code to test.

from ultralytics import YOLO

import numpy

model = YOLO(“yolov8m.pt”)

detection_output = model.predict(source= “C:/Users/wests/download/backend/test-image.png”, conf=0.25,save=True)

print(detection_output)

print(detection_output[0].numpy())

What am I doing wrong?

3

Hey there @MotoManWest welcome! This is a recent problem and I actually pinned it as an Issue on the GitHub repo as we were seeing many users encountering the same thing. Check out the write up there, but ultimately the best thing to do for now is downgrade your PyTorch install (see the “fixes” section on the Issue).