Udacity NDs exercises from (almost) scratch and redone for different frameworks, for practice
This repository contains several projects that are proposed along different lessons in the following ND:
- Computer Vision ND
- Deep Learning ND
- Natural Language ND
- AIND
- AWS Machine Learning Foundations Course
These are not the projects to be delivered for the courses, but the exercises that are being proposed along the way instead, at least most of the time. The objective is to bring the original exercise, solve it, and then - time allowed (ah! time...) - they will be translated to other frameworks, as Keras, TF or PyTorch. Ocassionally, they can be used too for new versions of these networks.
The collections of projects will start small, easy and simple and hopefully will grow as my experience and familiarity with them do it, too.
Each project will include the requirements file needed to reproduce the results, and if no data is present (as some data files are quite large), I will try to link them.
This is the basic exercise in which Andrew Trask present the LSMT to process reviews and determine if they convey a good review, or a bad review. It is presented in the Deep Learning ND and as a supplementary material in the CVND.
A basic and extensible package using basic Python 3.7 Anaconda installation (or just Python 3.7 plus Matplotlib) to demonstrate the creation of a package and following OOP principles. The package creates a general probabilistic distribution and two different base distribution have been defined:
- Gaussian: class It describes a Gaussian distribution by mean and standard deviation
- Binomial: class It describes a Binomial distribution by the probability of the binomial, and the number of elements in the population.