Dec 26, 2017Softmax is used in multi-class classification task. The Cross-Entropy enables to express the loss of the classifier. In this post, we will go through the mathematics to calculate the gradient of the cross-entropy loss function with respect to the weights.
Dec 6, 2017A tensor mask can be very useful when developing a Deep Learning Model. It can be used to turn on/off activations deterministically or randomly.
Nov 15, 2017A step-by-step tutorial on how to install Tensorflow for Python 3.5 with GPU support. The TF library is built from the source, and enables better system compatibility and better performance.
Mar 20, 2017I show how to implement a Custom Metric fnction in Keras, in particular the coefficient of determination. The coefficient of determination is particularly useful if we want to compare a model to a basic classification model based on the frequency of occurence of the classes in the training set.
Nov 29, 2016Here is a step by step on how to setup/install a NVidia GPU card on a Linux machine as a secondary graphic card for Deep Learning computation. That includes some tricks during the installation of the driver to prevent some known issues such as infinite login loop.
Nov 15, 2016This is a short post that shows a starting code to save and reload models created with Tensorflow. The model includes the weights and biases.
Aug 19, 2016In this post, we will identify customers segments using data collected from customers of a wholesale distributor in Lisbon (Portugal). The dataset includes the various customers annual spending amounts (reported in monetary units) of diverse product categories for internal structure.
Aug 19, 2016Boxplots is a an informative way to display the distribution of data. in this post, we will will go through the steps to create boxplots with seaborn.
Aug 1, 2016We will build a model that predicts how likely a student is to pass their high school final exam. The model must be effective while using the least amount of computation costs.
Jul 15, 2016In this project, we analyze the prices of homes in suburbs of Boston. We build a predictive model and train/test it on collected data. The performance of the model is then evaluated. The model can then be used to make certain predictions about a home — in particular, its monetary value. This model would prove to be invaluable for someone like a real estate agent who could make use of such information on a daily basis.
