TronicLT/notebooks

General Jupyter notebooks

Notebooks

General Jupyter notebooks

List of notebooks

1. UCF 101 - Action Bank: The notebook quantifies Action Bank features for action recognition using the UCF101 dataset.
2. Supervised dimensionality reduction: Notebook explains and implements MDA/LDA for supervised dimensionality reduction.
3. AlexNet conversion: This notebook example demonstrates how to convert a network from Caffe's Model Zoo for use with uml deep learning library (developed during master's work). AlexNet model trained for LSVRC2012. The notebook create a set of uml layers corresponding to the Caffe model specification (prototxt), then copy the parameters from the caffemodel file into our model. The conversion is verified using activation maximization.
4. Fraud Detection: Uses XGBoost trees, Random forest and Neural networks to classify between fraudulent and normal transactions that occured in two days. The dataset is highly unbalanced.