8Opt/yolo11-human-nonhuman-cls

YOLOv11n Human vs Non-Human Classification

This is a fine-tuned YOLOv11s-cls model for binary classification.

📊 Performance

• Test Accuracy: 98.24% (836/851 correctly classified)
• Dataset: Human faces vs. Non-human (Statues, Art, Anime, Gaming)
• Experiment ID: yolo11s_cls_260104_086

🚀 Model Usage

1. Using Ultralytics (Python)

The easiest way to use the trained PyTorch model:

from ultralytics import YOLO

# Load the model
model = YOLO('path/to/best.pt')

# Predict on an image
results = model('image.jpg')

# Process results
for result in results:
    probs = result.probs  # Probs object for classification outputs
    print(f"Top-1 class: {result.names[probs.top1]}")
    print(f"Confidence: {probs.top1conf:.2f}")

2. Using YOLO CLI

yolo classify predict model=path/to/best.pt source='image.jpg'

3. Using ONNX Runtime

For production or edge deployment without PyTorch:

import onnxruntime as ort
import numpy as np
import cv2

# Initialize session
session = ort.InferenceSession("model.onnx")

# Preprocess (Resize to 224x224, normalize [0, 1], CHW format)
img = cv2.imread("image.jpg")
img = cv2.resize(cv2.cvtColor(img, cv2.COLOR_BGR2RGB), (224, 224))
img = img.astype(np.float32) / 255.0
img = img.transpose(2, 0, 1)[np.newaxis, :]  # Add batch dim

# Run Inference
outputs = session.run(None, {"images": img})
predicted_idx = np.argmax(outputs[0]) 
print(f"Predicted Class Index: {predicted_idx}")

📈 Baseline Metrics (04/01/2026)

• Model: YOLOv11s-cls
• Input Size: 224x224
• Test Accuracy: 98.24%