🦋 Butterfly Classifier (PyTorch, Transfer Learning)

Classifies 75 butterfly species from images using a pretrained VGG16 with a small task-specific head.
Validation accuracy ≈ 90% on a held-out split.

Live demo: https://michaelmwb-butterfly-classifier.hf.space/

---

Repo contents

• train.py — training script (saves model.pth + stats.txt)
• ImageDataset.py — CSV-driven dataset & transforms
• categories.py — class names (CATEGORIES = [...]) in training order
• makeplots.py — turns stats.txt → learning.png (loss + accuracy)
• predict.py — batch inference (writes predictions.csv)
• app.py — small Gradio UI (upload image → top-k predictions)
• requirements.txt — Python deps
• learning.png — example learning curves (generated)
• predictions.csv — predictions of the model

Note on data: Training/validation/test images are not stored in this repo.
Download the data pack ➜ here

---

🛠 Usage Once the Requirements are Installed, Run:

python3 train.py   # May take a long time depending on CPU

⚡For NVIDIA GPU Users: To enable CUDA acceleration (recommended for training), install PyTorch with GPU support:

python3 -m pip install --index-url https://download.pytorch.org/whl/cu121 torch==2.5.1+cu121 torchvision==0.20.1+cu121 torchaudio==2.5.1+cu121

To Generate Learning Curves, Run:

python makeplots.py # Creates learning.png

To Run Batch Inference on Images:

python predict.py # Writes predictions labels to predictions.csv