yolov5 non_maxima_supression() keeps giving the error: 'Detections' object has no attribute 'device'

[MariaAhmed01]

I've been trying to run this code for yolov5 on a traffic police video but I can't figure out why it keeps giving me this error- even after specifying the device manually, it continues to give me this error.
This is the error:

and this is the code snippet:

# inference for videos
video_path = 'D:\\DL_A3\\traffic_Islamabad_Police.mp4'
video = cv2.VideoCapture(video_path)
ret, frame = video.read()

result = model(frame)

detections = result.pandas().xyxy[0]
detections = detections[['xmin', 'ymin', 'xmax', 'ymax', 'confidence']].to_numpy()

#convert the frame to tensor
img = torch.from_numpy(detections).to('cuda' if torch.cuda.is_available() else 'cpu')
img = img.float()/255.0

#manually specifying a device for the tensor
img_device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
img = img.to(img_device)

# Apply non-maximum suppression

detections = non_max_suppression(result, 0.25, 0.45, None, False, False, labels=(), max_det=1000)
frame = draw_bounding_boxes(frame, detections)

# Display the processed frame
cv2.imshow('Video Inference', frame)