Bears Classifier is a demo project showcasing a bear image classifier trained on the ResNet18 dataset using fastai. Upload photos of bears and let our model identify whether it's a grizzly, black, or polar bear. This demo is built to run smoothly in your browser and is deployed with GitHub Pages.
I kept things simple, and crafted this bear classifier with:
- JavaScript: For client-side functionality
- HTML: Structuring our web content
- CSS: Styling to make the interface clean and user-friendly
- FastAI: For training the deep learning model with minimal code complexity, chosen for its abstractions for training and fine-tuning neural networks.
- ResNet18: Deep residual learning model known for its efficiency and performance in image classification tasks.
- Hugging Face: To deploy the model and make it accessible via a web interface, we used Hugging Face Spaces.
- Data Collection: leveraged Kaggle's bear image dataset, which includes labeled images of grizzly, black, and polar bears.
- Model Training: Using FastAI's high-level API, we trained a ResNet18 model to classify bear images, employing transfer learning techniques to speed up the process and improve accuracy.
- Deployment: The trained model is deployed on Hugging Face Spaces, ensuring easy integration and real-time classification in the browser.
- 🐻 Upload images to classify different bear species
- 🔍 Accurate bear identification using ResNet18 and fastai
- 🌐 Deployed using GitHub Pages for easy access
- 📱 Responsive design to ensure great user experience on all devices
Want to run this demo locally? Here's how:
- Clone this repository:
git clone https://github.com/manonja/bears-classifier.git
- Navigate to the project repository:
cd bears-classifier
Have ideas to improve the classifier? We welcome your contributions! Whether it's fixing a bug or adding new features, feel free to fork the repository, create a feature branch, and send us a pull request!
For the training model, you can visit my notebook here
If you enjoy using Bears Classifier, share it with your friends and fellow bear enthusiasts!
This project is licensed under the Apache 2 License - see the LICENSE.md file for details.