train.py

# Copyright 2025 Robotics Group of the University of León (ULE)

# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at

#     http://www.apache.org/licenses/LICENSE-2.0

# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import os
from pathlib import Path
from ultralytics import YOLO


DATASET_NAME = "veridis"


def count_images(directory):
    """Count the number of JPG image files in a directory.

    Args:
        directory: Path to the directory containing images.

    Returns:
        Integer count of JPG files found in the directory.
    """
    if not os.path.exists(directory):
        return 0
    
    count = 0
    for file in os.listdir(directory):
        if file.lower().endswith(".jpg"):
            count += 1
    
    return count


def find_dataset_yaml(current_dir, dataset_name):
    """Find data.yml file in a specific dataset directory.

    Args:
        current_dir: Current directory to search from.
        dataset_name: Specific dataset name to use.

    Returns:
        Path to the data.yml file or None if not found.
    """
    dataset_path = current_dir / dataset_name
    yaml_path = dataset_path / "data.yml"
    
    if yaml_path.exists():
        print(f"Found dataset: {dataset_name}")
        return str(yaml_path), dataset_path
    
    return None, None


def main():
    """Main training function that initializes and trains the YOLO model.

    Counts dataset images, prints statistics, and executes model training
    with predefined hyperparameters.
    """
    current_dir = Path(__file__).parent
    os.chdir(current_dir)
    
    print(f"Training dataset: {DATASET_NAME}")
    yaml_path, dataset_root = find_dataset_yaml(current_dir, DATASET_NAME)
    
    if yaml_path is None:
        print(f"ERROR: Dataset '{DATASET_NAME}' not found or missing data.yml")
        print("\nAvailable datasets: veridis, beet_augmented, beet, corn_augmented, corn")
        return
    
    train_images = count_images(dataset_root / "train" / "images")
    val_images = count_images(dataset_root / "val" / "images")
    test_images = count_images(dataset_root / "test" / "images")
    
    print("\n=== Dataset Statistics ===")
    print(f"Dataset path: {dataset_root}")
    print(f"Config file: {yaml_path}")
    print(f"Training images: {train_images}")
    print(f"Validation images: {val_images}")
    print(f"Test images: {test_images}")
    print(f"Total images: {train_images + val_images + test_images}")
    print("=========================\n")
    
    print("Initializing YOLO model...")
    model = YOLO("yolo11n.pt")
    
    print("Starting training...\n")
    results = model.train(
        data=yaml_path,
        epochs=10,
        imgsz=640,
        batch=16,
        name=f"{dataset_root.name}_detection",
        project="runs/train",
        verbose=True
    )
    
    print("\n=== Training Complete ===")
    print(f"Results saved to: {results.save_dir}")
    
    print("\n=== Training Metrics ===")
    if hasattr(results, "results_dict"):
        metrics = results.results_dict
        for key, value in metrics.items():
            print(f"{key}: {value}")
    
    metrics = model.val()
    
    print("\n=== Validation Results ===")
    print(f"mAP50: {metrics.box.map50:.4f}")
    print(f"mAP50-95: {metrics.box.map:.4f}")
    print(f"Precision: {metrics.box.mp:.4f}")
    print(f"Recall: {metrics.box.mr:.4f}")
    
    if hasattr(metrics.box, "maps"):
        print("\nPer-class mAP50-95:")
        class_names = metrics.names
        for i, map_value in enumerate(metrics.box.maps):
            print(f"  {class_names[i]}: {map_value:.4f}")


if __name__ == "__main__":
    main()