PostGraduated-DataScience-CapstoneProject

Lucas Martínez personal Capstone Project for the UB Data Science Postgraduated course 2021-2022. http://www.ub.edu/datascience/postgraduate/

Notebook with the developed tests is available on Colab_Notebook/ folder. The notebook is configured to be run in Google Colab. Persistence is ensured via the google-drive asociated with the Collab user.

Quick user guide

Available CNN models

3 models are available to perform train, re-train or a posterior evaluation.

Set up mode01=True for a self-CNN model
Set up mode11=True for a ResNet based CNN model
Set up mode11=True for a DenseNet based CNN model

Training and Evaluating modes

When training:

Retrain: for a training from scrath set up retrain=False, to resume a previous training set up retrain=True
Epochs: set up num_epochs, to the desired number of epochs to train or retrain when evaluating:
Evaluation: Set up evaluating=True to make an evaluation of a previous training.

Report

Report will be available at: https://lumaro77.github.io/UB-DataScience-CapstoneProject/

Presentation

Available at presentacion and youtube

Barcelona, June 2022.