EnYa32's picture
Update README.md
516d97a verified
|
Raw History Blame Contribute Delete
1.07 kB
---
title: FashionMNISTClassification
emoji: 🧥
colorFrom: red
colorTo: red
sdk: streamlit
app_file: src/streamlit_app.py
pinned: false
short_description: CNN image classifier for Fashion-MNIST, 10 classes
license: mit
---
# 🧥 Fashion-MNIST CNN Classifier (Streamlit)
This Space demonstrates a Convolutional Neural Network (CNN) trained on the **Fashion-MNIST** dataset.
The model classifies grayscale clothing images into **10 classes** (e.g., T-shirt, Trouser, Sneaker, Bag).
## Model Performance
- Test Accuracy: **~92%**
## How it works
1. Upload an image (PNG/JPG).
2. The app converts it to **28×28 grayscale** and normalizes pixel values to **[0, 1]**.
3. The CNN predicts the class and shows probabilities.
> Note: The model was trained on Fashion-MNIST images (small grayscale). Real photos may reduce accuracy.
## Files
- `app.py` → Streamlit application
- `fashion_mnist_cnn.keras` → trained CNN model (must be in repo root)
- `requirements.txt` → dependencies
## Run locally
```bash
pip install -r requirements.txt
streamlit run app.py