👕 Classification of consumer goods

OPENCLASSROOMS - Data Scientist - Project 6

This repository contains notebooks for a machine learning project that predicts classes of consumer goods based on their picture and their description.

📊 Data

The dataset used for this project is 1050 articles defined by a description and a single picture.

📁 Files

• barbier_victor_1_notebook_112022.ipynb : Text and image machine learning models
• barbier_victor_2_bert_tuning_112022.ipynb : Fine-tuning of the BERT model on the dataset descriptions
• barbier_victor_3_cnn_tuning_112022.ipynb : Fine-tuning of the ResNet50 model on the dataset images
• barbier_victor_4_presentation_112022.pdf: Final presentation of the project

🛠️ Tools

• Python 3.x
• Jupyter Notebook
• NumPy
• Pandas
• Matplotlib / Seaborn
• Spacy
• Scikit-learn : KMeans, t-SNE
• Pytorch : BERT, ResNet50
• Tensorflow : Universal Sentence Encoder
• fastText
• ORB