How to calculate for each image while testing how much cpu utilization the yolo took to detect in this image?

Question

I am using yolo5. I want to estimate how much CPU utilization does yolo take to process a certain image. how can I do that?

I tried to use P-Sutil to calculate the CPU utilization however I can not get how much cpu utilization for each image it took

here is my code

# !pip install -U ultralytics
import torch
import time
import psutil
from keras.datasets import cifar10

# Load YOLOv5 model
# model = torch.hub.load('ultralytics/yolov5', 'yolov5s', pretrained=True)

# Load CIFAR-10 dataset and select 10 images
(_, _), (x_test, _) = cifar10.load_data()
x_test_subset = x_test[:10]  # Selecting first 10 images from the test set

# Loop through images
for i, img in enumerate(x_test_subset):
    # Measure CPU time before YOLOv5 inference
    start_cpu = psutil.cpu_times().user + psutil.cpu_times().system

    # Convert image to PIL format and perform YOLOv5 inference
    start_time = time.time()
    results = model("/content/yolo.jpg", size=640)  # You can adjust the size as needed
    end_time = time.time()

    # Measure CPU time after YOLOv5 inference
    end_cpu = psutil.cpu_times().user + psutil.cpu_times().system

    # Calculate CPU utilization for YOLOv5 inference
    cpu_utilization_percentage = ((end_cpu - start_cpu) / psutil.cpu_count()) / (end_time - start_time) * 100
    
    # Count objects detected
    num_objects = len(results.xyxy[0])

    # Print results for the current image
    print("Image {}: Objects detected: {} - YOLO CPU utilization (%): {:.2f}".format(i+1, num_objects, cpu_utilization_percentage))
`