Common is_coco logic betwen train.py and val.py

[glenn-jocher]

Resolves #3644

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Update to COCO dataset detection logic in training script.

📊 Key Changes

• Adjusted the method to identify if the COCO dataset is being used.
• Previously, checked if the data configuration ended with 'coco.yaml' and had 80 classes.
• Now, checks if the val_path is a string that ends with 'coco/val2017.txt'.

🎯 Purpose & Impact

• 🎨 Ensures more accurate detection of COCO dataset use by examining validation path.
• ⚙️ Potentially impacts users training models with COCO dataset, improving reliability for model setup.
• 🚀 Users can expect a more robust way of identifying the usage of COCO dataset, leading to fewer errors and assumptions during training.