PyTorch-Linear-DNN-Fundamentals

This project serves as a step-by-step introduction to the PyTorch deep learning workflow, demonstrating how a simple neural network learns the pattern of a straight line (y = weight * x + bias).

It implements the full cycle of data preparation, model construction, training, evaluation, and saving/loading, following a device-agnostic approach (works on both CPU and GPU).

Core Concepts Explained

A deep learning workflow typically involves these steps:

1. Data Preparation: Create and split the dataset (training, testing).
2. Model Building: Define the neural network architecture.
3. Loss Function & Optimizer: Define how the model measures errors (Loss) and how it updates its parameters to reduce those errors (Optimizer).
4. Training Loop: The iterative process of Forward Pass, Loss Calculation, Backpropagation, and Optimizer Step.
5. Inference: Using the trained model to make predictions.
6. Saving/Loading: Persisting the learned parameters.

Files

• linear_regression_pytorch.py: The main, self-contained Python script implementing the entire workflow.
• requirements.txt: Lists necessary Python libraries.

Setup and Running

1. Prerequisites

You need Python (3.8+) and the following libraries:

torch
matplotlib

2. Run Locally (Recommended)

1. Clone the repository:

   git clone https://github.com/Rahmat-ML/PyTorch-Linear-DNN-Fundamentals
   cd PyTorch-Linear-DNN-Fundamentals
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Run the script:

   python linear_regression_pytorch.py
   

   The script will print training progress, model parameters, and save plots/the final model state.