XGBoost and CNN model implementation with Ray Framework:

Description:

This project demonstrates the implementation of distributed machine learning models using the Ray framework. It includes two distinct models:

• XGBoost: A classification model trained on the Breast Cancer dataset.
• CNN: An image classification model trained on the CIFAR-10 dataset.

The focus of this project is to evaluate the efficiency and scalability of the Ray framework in reducing execution time while maintaining robustness through fault tolerance and hyperparameter optimization.

Requirements

Hardware requirements:

• For XGBoost:

  • 4 CPUs
  • 16GB RAM
  • 64GB Memory

• For CNN:

  • 2 CPUs
  • 1 GPUs (minimum)
  • 25GB RAM for CPUs & 16GB RAM for GPU
  • 112GB Memory

Software requirements:

• Python 3.10.11
• Libraries:
  • ray
  • tensorflow
  • keras
  • xgboost
  • scikit-learn

How to Run: (Approach - 1)

1. Run the batch file that is present in the github repo named as 'Run.bat' file.
   • Navigate to directory where your batch file is located and replace the 'path_to_your_project_directory' to your path.

   cd path_to_your_project_directory

2. Run the batch file by adding the command 'run.bat'

   run.bat
   

How to Run: (Approach - 2)

You can directly double click on run.bat file and execution of python files will start successfully ensuring all the resources are available.