Project Overview
This project demonstrates how I built an image classification model to distinguish between cats and dogs using TensorFlow, Keras, and Streamlit.
The workflow covers the entire deep learning lifecycle:
- Data preprocessing with
ImageDataGenerator - Model building using Convolutional Neural Networks (CNNs)
- Training & evaluation on filtered Cats vs Dogs dataset
- Saving the model & label mapping for reproducibility
- Deployment as an interactive Streamlit app where users upload an image and instantly get predictions
Key Features
- ✅ End-to-end Deep Learning pipeline with TensorFlow
- ✅ Custom CNN trained on Cats vs Dogs dataset
- ✅ Model + class mapping stored (
cnn_model.h5 + class_indices.json) - ✅ Interactive web app built with Streamlit
- ✅ Real-time predictions with confidence scores
- ✅ SEO-ready write-up for portfolio visibility
Tech Stack
- Python 3.12
- TensorFlow / Keras
- NumPy, Pandas
- Matplotlib, Seaborn (for training analysis)
- Streamlit (for deployment)
Outcomes
The trained model achieves solid performance in classifying cats vs dogs and can be extended to other binary/multi-class image classification tasks.
This project shows my ability to:
- Work with real-world datasets
- Implement CNN architectures
- Train, evaluate, and fine-tune models
- Build practical AI apps for end-users
Live Demo & Code
