README.md

---
title: Keras Image Classifier
emoji: 🖼️
colorFrom: indigo
colorTo: blue
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
pinned: false
---


# Image Classifier — Keras/TensorFlow (Hugging Face Space)

A dead-simple image classification app you can deploy in minutes.

## How it works

- If `model.h5` exists in the repository root, the app loads **your custom Keras model**.
  - Optionally add `labels.txt` (one class name per line) to show readable labels.
  - Input is resized to **224×224**. Adjust `TARGET_SIZE` in `app.py` if your model expects a different size.
- If no `model.h5` is found, it falls back to **MobileNetV2 (ImageNet)**.

## Run locally

```bash
pip install -r requirements.txt
python app.py
```

Then open the local URL printed by Gradio.

## Deploy to Hugging Face Spaces

1. Create a new **Space** → **Gradio** (Python).
2. Upload these files: `app.py`, `requirements.txt`, `README.md`.
3. (Optional) Upload your `model.h5` and `labels.txt` to use your own model.
4. The Space will build and auto-start.

## Using your notebook's model

If your notebook trained a model, export it:

```python
model.save("model.h5")
# Optional labels file (one per line)
with open("labels.txt", "w") as f:
    f.write("\n".join(class_names))
```

Commit both files to the Space.