https://github.com/piyush-kumar-ghosh/digit-predictor
Streamlit web app that uses a CNN model to predict handwritten digits using the MNIST dataset.
https://github.com/piyush-kumar-ghosh/digit-predictor
cnn computer-vision deep-learning image-classification machine-learning mnist portfolio-project streamlit-webapp tensorflow
Last synced: 2 months ago
JSON representation
Streamlit web app that uses a CNN model to predict handwritten digits using the MNIST dataset.
- Host: GitHub
- URL: https://github.com/piyush-kumar-ghosh/digit-predictor
- Owner: Piyush-Kumar-Ghosh
- Created: 2025-04-04T12:41:44.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-04-04T13:06:15.000Z (over 1 year ago)
- Last Synced: 2025-04-04T13:40:49.230Z (over 1 year ago)
- Topics: cnn, computer-vision, deep-learning, image-classification, machine-learning, mnist, portfolio-project, streamlit-webapp, tensorflow
- Language: Python
- Homepage: https://digit-predictor-hqfeqrpxxs5vvkgefv7ltp.streamlit.app/
- Size: 0 Bytes
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# 🧠 MNIST Digit Recognizer - Streamlit App
This is a simple and interactive web app that recognizes handwritten digits using a Convolutional Neural Network (CNN) trained on the MNIST dataset. Built with [Streamlit](https://streamlit.io), this app lets users draw digits directly in the browser and get real-time predictions.
---
## 🚀 Live Demo
🔗 [Try the app on Streamlit](https://digit-predictor-hqfeqrpxxs5vvkgefv7ltp.streamlit.app/)
---
## 🖼️ App Features
- Draw a digit (0-9) in the canvas
- Live prediction with confidence scores
- Built using TensorFlow + Streamlit
---
## 🧠 Model Info
- Trained on the MNIST dataset
- 2 Convolutional layers + Dense layers
- Accuracy: ~98% on test set
- Saved as: `mnist_digit_model.keras`
---
## 🛠️ How to Run Locally
```bash
git clone https://github.com/Piyush-Kumar-Ghosh/digit-predictor.git
cd digit-predictor
pip install -r requirements.txt
streamlit run app.py