Optimizing XARM7 with Mini PC AMD for Culinary Image Analys

This script is designed to train a YOLO (You Only Look Once) model using the ultralytics library. YOLO is a popular deep learning model for real-time object detection and segmentation. The provided script performs several tasks, including loading training and validation data, training the model, validating it, and saving the best model checkpoint.

Breakdown of the Script
Import Libraries

python
Copy code
from ultralytics import YOLO
import os
Define the train_model Function

Paths for training and validation images and labels are defined.
It checks the number of images and labels and ensures they match.
Initializes the YOLO model.
Trains the model using the specified configuration.
Validates the trained model.
Saves the best model checkpoint.
python
Copy code
def train_model():
train_images_path = 'D:/project/competition/AMD/Xarm7_making_sandwich/Media/Ingredients.v11i.yolov8/train/images'
train_labels_path = 'D:/project/competition/AMD/Xarm7_making_sandwich/Media/Ingredients.v11i.yolov8/train/labels'
val_images_path = 'D:/project/competition/AMD/Xarm7_making_sandwich/Media/Ingredients.v11i.yolov8/valid/images'
val_labels_path = 'D:/project/competition/AMD/Xarm7_making_sandwich/Media/Ingredients.v11i.yolov8/valid/labels'

print(f"Training images: {len(os.listdir(train_images_path))}")
print(f"Training labels: {len(os.listdir(train_labels_path))}")
print(f"Validation images: {len(os.listdir(val_images_path))}")
print(f"Validation labels: {len(os.listdir(val_labels_path))}")

assert len(os.listdir(train_images_path)) == len(os.listdir(train_labels_path)), "Mismatch between train images and labels"
assert len(os.listdir(val_images_path)) == len(os.listdir(val_labels_path)), "Mismatch between validation images and labels"

model = YOLO('yolov8n-seg.yaml')

print("Starting training...")
model.train(data='D:/project/competition/AMD/Xarm7_making_sandwich/Media/Ingredients.v11i.yolov8/data.yaml', epochs=100, imgsz=640)
print("Training complete.")

print("Validating model...")
results = model.val()
print(results)

if model.ckpt is None:
print("Error: Model checkpoint is None. Skipping model saving.")
else:
print("Saving model checkpoint...")
model.save('best.pt')
print("Model saved successfully.")
Execute the train_model Function

The script will run the train_model function if it is executed as the main module.
python
Copy code
if __name__ == '__main__':
train_model()
How to Use the Script
Set Up Your Environment

Ensure you have Python installed.
Install the ultralytics library, which includes the YOLO implementation:
bash
Copy code
pip install ultralytics
Prepare Your Data

Organize your training and validation images and their corresponding labels in the specified directories.
Ensure your dataset configuration file (data.yaml) is properly set up.
Run the Script

Execute the script in your terminal or command prompt:
bash
Copy code
python your_script_name.py
Key Points
Paths: The paths to images and labels need to be correctly set. Ensure the directories exist and contain the expected files.
Assertions: The script checks that the number of images matches the number of labels for both training and validation datasets.
Model Configuration: The YOLO model is initialized using a configuration file (yolov8n-seg.yaml). Make sure this file exists and is correctly configured for your use case.
Training and Validation: The model is trained for 100 epochs with an image size of 640. Adjust these parameters as needed.
Saving the Model: The best model checkpoint is saved to a file named best.pt.
This script automates the training process of a YOLO model for object detection or segmentation tasks, provided that the data is well-prepared and organized according to the paths and configuration files specified in the script.