from_scratch

As the name suggests, this repo is all about building things with pure math—from the ground up. Whether it’s a neural network in Python or a regression model in C++, I’m diving into every equation, every line of code, and having fun on my way to becoming invincible <3.

---

📂 Repository Structure

harshitjais14-from_scratch/
├── README.md             # This file
├── MLinCPP/              # Machine learning in C++ implementations
│   ├── dataset.txt       # Example datasets
│   ├── generatedataset.ipynb  # Notebook for generating synthetic data
│   ├── input.txt         # Sample input format
│   ├── output.txt        # Expected output format
│   ├── polynomialregression.cpp  # C++ implementation of polynomial regression
│   └── .vscode/          # Editor settings for VS Code
│       ├── settings.json
│       └── tasks.json
└── Python/               # Neural network and graph explorations in Python
    ├── andrej_nn/        # Micrograd-based neural network from Andrej Karpathy
    │   ├── building_micrograd.ipynb  # Building your own autograd engine
    │   ├── makemore.ipynb # Character-level probabilistic model
    │   ├── makemore2.ipynb # Character level probabilistic model using larget context window 
    │   └── names.txt      # Dataset for name generation
    ├── LangGraph/        # Experimenting with graph neural networks
    │   ├── helloworldgraph.ipynb   # Graph basics tutorial
    │   ├── graph2.ipynb            # Second GNN example
    │   └── graph3.ipynb            # Advanced GNN exploration
    ├── NNs/              # Neural networks from scratch
    │   ├── nns_scratch.ipynb   # Basic feedforward neural network
    │   └── cnns_scratch.ipynb  # Convolutional neural network implementation
    └── .vscode/          # Editor settings for VS Code
        └── settings.json

---