Hummingbird vs XGBoost: What are the differences?

### **Key Differences between Hummingbird and XGBoost**
Hummingbird and XGBoost are both popular machine learning tools, but they have distinct differences that make each unique. 
1. **Execution Mode:** Hummingbird is designed to optimize the deployment of trained models on various platforms, using tools like Databricks or PySpark for large-scale distributed computing, while XGBoost focuses on training models efficiently within a single machine.
2. **Library Support:** Hummingbird supports a wider range of machine learning libraries, allowing users to convert models from popular frameworks like TensorFlow and PyTorch to boosters like LightGBM and CatBoost, whereas XGBoost focuses primarily on boosting algorithms.
3. **Parallelism:** XGBoost utilizes parallel computation during training to improve speed, while Hummingbird primarily leverages parallelism during deployment to scale across multiple machines.
4. **Integration with Existing Systems:** Hummingbird aims to integrate seamlessly with existing ML pipelines and systems, enabling easy model deployment, while XGBoost is more suited for standalone model training and inference tasks.
5. **Model Size and Portability:** Hummingbird focuses on minimizing the size of the deployed model to enhance portability and reduce memory footprint, while XGBoost may have larger model sizes due to its training methodology and focus on accuracy.
6. **Flexibility:** While XGBoost offers flexibility in tuning hyperparameters and ensemble methods, Hummingbird streamlines the conversion and deployment process for a faster time-to-production for machine learning models.

In Summary, Hummingbird and XGBoost differ in execution mode, library support, parallelism, integration with existing systems, model size and portability, and flexibility, catering to different needs in the machine learning landscape.