deeplearning-lab-ktu

Getting Started with Deep Learning Labs

Quick Start Guide

Welcome to the Deep Learning Lab collection! This guide will help you get started quickly.

Prerequisites

Installation

Step 1: Clone or Download

If you haven’t already, navigate to the dl-lab directory:

cd dl-lab

# Create virtual environment
python3 -m venv venv

# Activate virtual environment
# On macOS/Linux:
source venv/bin/activate

# On Windows:
venv\Scripts\activate

Step 3: Install Dependencies

# Install all required packages
pip install -r requirements.txt

This will install:

Step 4: Verify Installation

python3 -c "import torch; print(f'PyTorch: {torch.__version__}')"
python3 -c "import cv2; print(f'OpenCV: {cv2.__version__}')"

Lab Structure

dl-lab/
├── README.md                          # Main overview
├── GETTING_STARTED.md                 # This file
├── requirements.txt                   # All dependencies
├── run_all_labs.sh                    # Run all labs script
│
├── lab_01_image_processing/             # Basic image operations
│   ├── README.md
│   ├── image_processing.py
│   └── outputs/
│
├── lab_02_cifar10_classifiers/          # Classification comparison
│   ├── README.md
│   ├── cifar10_classifiers.py
│   └── outputs/
│
├── lab_03_batchnorm_dropout/            # Regularization study
│   ├── README.md
│   ├── batchnorm_dropout_study.py
│   └── outputs/
│
├── lab_04_labeling_tools/               # Annotation guide
│   └── README.md
│
├── lab_05_segmentation/                 # Image segmentation
│   └── README.md
│
├── lab_06_object_detection/             # Object detection
│   └── README.md
│
├── lab_07_image_captioning/             # Image to text
│   └── README.md
│
├── lab_08_chatbot/                      # Conversational AI
│   └── README.md
│
├── lab_09_time_series/                  # Time series forecasting
│   └── README.md
│
└── lab_10_seq2seq/                     # Sequence to sequence
    └── README.md

Running Individual Labs

Lab 1: Image Processing

cd lab_01_image_processing
python3 image_processing.py

Time: ~30 seconds
Output: 6 visualization images in outputs/

Lab 2: CIFAR-10 Classifiers

cd lab_02_cifar10_classifiers
python3 cifar10_classifiers.py

Time: ~3-4 minutes
Output: Comparison plots and confusion matrices

Lab 3: Batch Normalization & Dropout

cd lab_03_batchnorm_dropout
python3 batchnorm_dropout_study.py

Time: ~3-4 minutes
Output: Training curves and performance comparison

Labs 4-10

These labs are primarily educational with comprehensive README files:

Each lab’s README contains:

Running All Labs

To run all executable labs sequentially:

# Make script executable (first time only)
chmod +x run_all_labs.sh

# Run all labs
./run_all_labs.sh

What to Expect

Lab 1 Output:

SUCCESS: Sample image created
SUCCESS: Histogram equalization complete
SUCCESS: Thresholding complete
SUCCESS: Edge detection complete
SUCCESS: Data augmentation complete
SUCCESS: Morphological operations complete

Lab 2 Output:

Training KNN...
  SUCCESS: Test Accuracy: 38.20%
Training SVM...
  SUCCESS: Test Accuracy: 42.50%
Training Neural Network...
  SUCCESS: Test Accuracy: 52.80%

Lab 3 Output:

Training Baseline...
  SUCCESS: Final validation accuracy: 45.30%
Training Batch Normalization...
  SUCCESS: Final validation accuracy: 52.60%
Training Dropout...
  SUCCESS: Final validation accuracy: 49.20%
Training BatchNorm + Dropout...
  SUCCESS: Final validation accuracy: 54.70%

Learning Path

Beginner Path:

  1. Lab 1: Image Processing Basics
  2. Lab 2: Classification Algorithms
  3. Lab 4: Data Annotation
  4. Lab 3: Regularization Techniques

Intermediate Path:

  1. Lab 5: Image Segmentation
  2. Lab 6: Object Detection
  3. Lab 9: Time Series Forecasting

Advanced Path:

  1. Lab 7: Image Captioning
  2. Lab 8: Chatbot Development
  3. Lab 10: Seq2Seq Learning

Troubleshooting

Issue: Import Errors

# Reinstall packages
pip install --upgrade -r requirements.txt

Issue: CUDA Out of Memory

# In the code, change device to CPU
device = torch.device('cpu')

Issue: Slow Training

# Use smaller datasets or reduce epochs
# Most labs have use_subset=True option

Issue: Permission Denied (run_all_labs.sh)

chmod +x run_all_labs.sh

Tips for Success

  1. Start with Lab 1: It’s the easiest and fastest
  2. Read the READMEs: Each lab has detailed documentation
  3. Experiment: Modify parameters and see what happens
  4. Use GPU: If available, training will be much faster
  5. Take Notes: Document your observations
  6. Ask Questions: Use the resources provided in each lab

Additional Resources

Online Courses:

Books:

Documentation:

Contributing

Found an issue or want to improve a lab?

  1. Document the issue or improvement
  2. Test your changes
  3. Share your modifications

Lab Completion Checklist

Next Steps

After completing these labs:

  1. Build Your Own Project: Apply what you learned
  2. Kaggle Competitions: Test your skills
  3. Research Papers: Read latest developments
  4. Open Source: Contribute to projects
  5. Share Knowledge: Teach others

Support

For questions or issues:

Congratulations!

You’re now ready to start your deep learning journey. Begin with Lab 1 and work your way through. Each lab builds on previous concepts, so take your time and enjoy learning!

Happy Learning!

Last Updated: March 2026