---
title: "lepus-classifier"
description: "A CNN research project exploring optimal image classification architectures for small datasets, demonstrating that data quantity remains the fundamental bottleneck for deep learning performance."
kind: research
status: archived
language: "Jupyter Notebook"
stars: 2
githubUrl: https://github.com/urmzd/lepus-classifier
pushedAt: 2026-03-16
tags: ["bash", "cnn", "computer-vision", "deep-learning", "docker", "github-actions", "image-classification", "machine-learning", "opencv-python", "pip", "python", "pytorch", "pytorch-lightning", "research", "small-dataset"]
---

# lepus-classifier

A CNN research project exploring optimal image classification architectures for small datasets, demonstrating that data quantity remains the fundamental bottleneck for deep learning performance.

- GitHub: https://github.com/urmzd/lepus-classifier

> **Archived.** This was a group project / final assessment for [CSCI 4155/6505 — Machine Learning](https://web.cs.dal.ca/~tt/CSCI415512/) at Dalhousie University, part of the [Artificial Intelligence & Intelligent Systems certificate](https://www.dal.ca/faculty/computerscience/undergraduate-programs/program-planning/certificates.html) (undergraduate/graduate mixed course). It is no longer actively maintained. Note: this project was built on PyTorch Lightning, which has since been superseded by [Lightning](https://github.com/Lightning-AI/pytorch-lightning).

<p align="center">
  <h1 align="center">Lepus Classifier</h1>
  <p align="center">
    A CNN-based image classifier that distinguishes between two rabbit species using only ~85 training images.
    <br /><br />
    <a href="https://github.com/urmzd/lepus-classifier/releases">Install</a>
    &middot;
    <a href="https://github.com/urmzd/lepus-classifier/issues">Report Bug</a>
    &middot;
    <a href="https://github.com/urmzd/lepus-classifier/blob/main/README.md">Research Report</a>
  </p>
</p>

<p align="center">
  <a href="https://github.com/urmzd/lepus-classifier/actions/workflows/ci.yml"><img src="https://github.com/urmzd/lepus-classifier/actions/workflows/ci.yml/badge.svg" alt="CI"></a>
</p>

<img src="./assets/mascot.webp" alt="mascot" width="200">

## Quick Start

1. Install dependencies:

```bash
pip install -r requirements.txt
```

2. Login to [Weights & Biases](https://wandb.ai) for experiment tracking:

```bash
wandb login
```

3. Train a model:

```python
from src.bootstrap import bootstrap

bootstrap()
```

Or define a custom architecture:

```python
from src.bootstrap import bootstrap
from src.data.model import BaseModel

class CustomModel(BaseModel):
    def __init__(self):
        super().__init__()
        self.layers = torch.nn.Sequential(...)
        super().__post_init__()

    def forward(self, x):
        return self.layers(x)

bootstrap(model=CustomModel())
```

See [notebooks/example_bootstrap.ipynb](notebooks/example_bootstrap.ipynb) for a Google Colab example.

## Project Structure

| Path | Description |
|------|-------------|
| [`src/`](src/) | Training pipeline, data loading, and model definitions |
| [`notebooks/`](notebooks/) | Experiment notebooks and usage examples |
| [`resources/data.csv`](resources/data.csv) | Dataset manifest (image URLs and labels) |
| [`docs/report-docs/lepus-classifier-report.pdf`](docs/report-docs/lepus-classifier-report.pdf) | Final research report |
| [`docs/poster.pdf`](docs/poster.pdf) | Project poster |
| [`docs/proposals/`](docs/proposals/) | Initial project proposals |

## Experiment Logs

Tracked with Weights & Biases: [csci-4155-rabbit-classifier](https://wandb.ai/csci-4155-rabbit-classifier)

## Agent Skill

```bash
npx skills add urmzd/lepus-classifier
```

## License

[Apache 2.0](LICENSE)