shoes-recommendation

Shoes recommendation based on image similarity using NNextDB technology

Install

To run this demo you need to install:

• Python 3.8.8 (https://www.python.org/downloads/)
• Poetry (https://python-poetry.org/docs/)

Dataset

1. Download shoes images dataset from the following link: https://www.kaggle.com/datasets/die9origephit/nike-adidas-and-converse-imaged
2. Unzip the file
3. Move it to the root dir of this repo
4. Rename folder to data

Feature vector extractor

If you want to fine tune the Resnet-18 model checkpoint you can use Resnet_Finetuning.ipynb notebook. Remember that Technical_Implementation.ipynb uses this checkpoint file.

Dependencies

1. From the root dir of this repo execute poetry shell. (this will activate the new virtual environment)
2. poetry install will automatically download dependencies.

Execute notebook

Now you are good to execute Technical_Implementation.ipynb demo notebook