This repository contains the jupyter notebooks that I have made while learning to explore data and do predictions with machine learning models.
Some of them are written in french.
The projects are the following (from most recent to less recent):
🐟 interactive image segmentation with photos of fish proof of concept: comparing performances of SAM (Meta) with RITM on dataset of pictures of fish. [📒 Notebook on nbviewer] [Link to repo].
🐕 dog breed image classification Classification of photos of dogs by breed using transfer learning with VGG16 and EfficientNet. This algorithm can be tested online in this HugginFace space. [📒 Notebook on nbviewer]
👥 stackoverflow tag suggestion: Creation of an algorithm of automatic tag suggestion based on text. Here I used some NLP machine learning libraries and known language embedding models. This algorithm can be tested online in this HugginFace space. [📒 Notebook 1 on nbviewer] [📒 Notebook 2 on nbviewer]
🛒 Olist client segmenation: As name states: finding an interesting segmentation of clients of an e-commerce website. [📒 Notebook 1] [📒 Notebook 2] [📒 Notebook 3]
🏢 SeattleEnergyBenchmarking: exploration of data from 2016 and prediction of energy use for non-residential buildings. [📒 Notebook 1 on nbviewer] [📒 Notebook 2 on nbviewer]
🥪 OpenFoodFacts: exploration of the Open Food Facts database. [📒 Notebook 1 on nbviewer] [📒 Notebook 2 on nbviewer]