Deep Learning Exploration with PyTorch

Hello to my first GitHub space! This project is a practical look at artificial intelligence. It mainly learns deep-learning ideas using PyTorch. The collection has two Jupyter Colab notebooks that show different aspects of deep-learning.

Notebooks Overview

1. Iris Dataset Exploration

• In this notebook (iris_dataset_classification.ipynb), I've implemented a simple 3-layer neural network for classification using the classic Iris dataset.
• Explored the impact of varying epochs and learning rates on the model's loss and accuracy.
• Visualized loss curves and accuracy trends for different hyperparameter settings.

2. Non-Linearity in Deep Learning

• The second notebook (non_linearity_demo.ipynb) demonstrates the importance of non-linearity in deep learning.
• Used the make_circles dataset for binary classification without non-linearity.
• Extended the same model for regression using a synthetic linear dataset.
• Introduced a ReLU activation function to showcase the power of non-linearity in the make_circles dataset.

How to Use

1. Open the respective notebook in Google Colab or Jupyter Notebook.
2. Run the cells sequentially to observe the results and visualizations.
3. Experiment with hyperparameters to gain insights into model behavior.

Future Directions

I plan to expand on this exploration by adding more notebooks that delve into advanced topics such as transfer learning, convolutional neural networks, and natural language processing. Your feedback and contributions are highly welcome!

License

Feel free to use and modify the code for your own learning and projects.

Acknowledgments

Special thanks to the PyTorch community for providing an amazing framework for deep learning experimentation. I'd like to express my gratitude to mrdbourke for the helper functions I've imported into my code.

Happy coding!