README.md

🐾 Fox Classifier

A simple ML project for classifying fox (and cat just) using PyTorch, FastAPI, and Streamlit.
The application runs via Docker Compose and consists of two services:

• API (FastAPI + Uvicorn) — performs predictions using the trained model.
• UI (Streamlit) — provides a user-friendly web interface to upload images and get results.

---

📂 Project Structure

project-root/
│── foxes_for_TA             # A folder with Fox images for network testing, try them all, I searched very diligently for them  :)
│── code/
│   ├── models/              # model training
│   ├── deployment/
│   │   ├── api/             # FastAPI service and best model storage
│   │   └── app/             # Streamlit service
│── docker-compose.yml       # runs both services
│── README.md

---

🚀 Getting Started

1. Build and run containers

docker compose up --build

After that the following services will be available:

• API: http://localhost:8000/docs (Swagger UI)
• Streamlit UI: http://localhost:8501

---

📖 Usage

Via UI

1. Open http://localhost:8501.
2. Upload an image (cat / fox).
3. Get a prediction from the model.

Via API

You can send a request directly:

curl -X POST "http://localhost:8000/predict/" \
  -F "file=@fox.jpg"

Response:

{
  "class": "fox
}

---

⚙️ Tech Stack

• Python 3.10
• PyTorch + Torchvision
• FastAPI
• Streamlit
• Docker & Docker Compose

Update NN

In root of repo run following command

python code/models/thinker_main.py 

---

🛠️ TODO

• Add model training inside a container
• Store prediction history
• Extend dataset and classes