Add TF and TFLite models to .dockerignore

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Enhanced .dockerignore to support additional machine learning model formats.

📊 Key Changes

• Added ignore rules for TensorFlow Lite models (*.tflite).
• Added ignore lines for Keras model files (*.h5).
• Included TensorFlow saved model directories (*_saved_model/) in the ignore list.
• Incorporated ignore patterns for TensorFlow protobuf files (*.pb).

🎯 Purpose & Impact

• Purpose: Streamline Docker builds by excluding unnecessary files, which can help reduce context size and consequently the build time and size of the images.
• Impact: Users working with TensorFlow and Keras models will benefit from cleaner Docker builds, leading to faster deployment and potentially lower resource usage.