- Vanilla CNN (baseline) - 71% acc cats_vs_dots_cnn.ipynb
- Vanilla CNN with data augmentation and dropout - 85% acc cats_vs_dots_cnn.ipynb
- Fine-tuned VGG19 model with data augmentation and dropout - 93% acc cats_vs_dots_VGG19.ipynb
- Fine-tuned ResNet50 model with data augmentation and dropout - 96% acc cats_vs_dots_ResNet50.ipynb
- Visualizing layer activations on a test image for fine-tuned VGG19 and ResNet50 cats_vs_dots_layer_activations.ipynb
- Filter visualization using gradient-ascent for fine-tuned ResNet50 cats_vs_dots_grad_ascent.ipynb
- Grad-CAM visualization for fine-tuned ResNet50 cats_vs_dots_grad_cam.ipynb
- Deep Dream
- Vanilla CNN - 91% acc.
- Convolutional variational autoencoder for image generation
- k-Nearest Neighbours classification
- Logistic Regression
- Logistic Regression using polynomial features
- SVM with linear kernel, binary classification, multi-class classification
- Decision Tree
- Gaussian Naive Bayes
- Random Forest
- k-Nearest Neighbours regression
- Linear Regression
- Ridge Regression (Regularization)
- Ridge Regression with feature scaling
- Polynomial Regression
- k-Means Clustering
- DBSCAN
- Spectral Clustering
- Markov Clustering (markov_clustering Python module)