CPU vs GPU Training Time Comparison with TensorFlow

This project demonstrates a comparison of training times for a ResNet50 model on the CIFAR-10 dataset using TensorFlow, running on both CPU and GPU.

---

Table of Contents

• Overview
• Technologies Used
• Setup and Requirements
• Running the Code

---

Overview

This demo is part of a learning module designed to provide hands-on experience in comparing computational performance on different hardware setups. Specifically, the goals are to:

• Train a ResNet50 model on the CIFAR-10 dataset.
• Compare the training times on a CPU and a GPU.
• Visualize the training time difference using a bar chart.
• Gain practical insights into optimizing machine learning workflows for various hardware configurations, a critical aspect of modern AI and ML development.

---

Technologies Used

• Programming Language: Python
• Deep Learning Framework: TensorFlow
• Architecture: ResNet50
• Visualization: Matplotlib
• Dataset: CIFAR-10
• Hardware: NVIDIA GPU (CUDA-enabled)

---

Setup and Requirements

Prerequisites

• Python 3.8 or later
• TensorFlow 2.0 or later
• Matplotlib
• NVIDIA GPU with CUDA support (if using a GPU)

Installation

1. Clone the repository:

   https://github.com/MohanKrishnaGR/cpu-vs-gpu-ResNet50-training-cifar10.git
   

2. Navigate to the project directory:

   cd cpu-vs-gpu-ResNet50-training-cifar10
   

3. Install required Python libraries:

   pip install 'tensorflow[and-cuda]' matplotlib
   

---

Running the Code

1. Check GPU availability:

   !nvidia-smi
   

2. Execute the Python script:

   python train_resnet_comparison.py
   

3. The script will:

• Train the ResNet-like model on the CIFAR-10 dataset using both CPU and GPU.
• Record the training time for each device.
• Display a bar chart comparing the training times.

---

End Note

Thank you for your interest in this demo! We welcome any feedback. Feel free to reach out to us.