# Img2VecCosSim-Django-Pytorch

**Repository Path**: dyjch/Img2VecCosSim-Django-Pytorch

## Basic Information

- **Project Name**: Img2VecCosSim-Django-Pytorch
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: GPL-3.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2021-02-20
- **Last Updated**: 2021-11-11

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Img2VecCosSim-Django-Pytorch
[![HitCount](http://hits.dwyl.com/MexsonFernandes/Img2VecCosSim-Django-Pytorch.svg)](http://hits.dwyl.com/MexsonFernandes/Img2VecCosSim-Django-Pytorch)

Extract a feature vector for any image and find the cosine similarity for comparison using Pytorch. I have used ResNet-18 to extract the feature vector of images. Finally a Django app is developed to input two images and to find the cosine similarity.

<a href='https://ko-fi.com/Y8Y31LBT4' target='_blank'><img height='36' style='border:0px;height:36px;' src='https://cdn.ko-fi.com/cdn/kofi3.png?v=2' border='0' alt='Buy Me a Coffee at ko-fi.com' /></a>

#### Packages:
  - Pytorch
  - Django 2.0
  
#### Credits:
  - Inspired from <a href="https://becominghuman.ai/extract-a-feature-vector-for-any-image-with-pytorch-9717561d1d4c">Medium post by Christian Safka</a>
  
#### How to start:
  * Clone repository
  
    `git clone https://github.com/MexsonFernandes/Img2VecCosSim-Django-Pytorch`
  * Change directory
  
    `cd Img2VecCosSim-Django-Pytorch`
  * Install virtual environment
  
    `pipenv install`
  * Install all dependencies
  
    `pipenv install -r requirements.txt` or `pip install -r requirements.txt`
    
  * Start django server
  
    `python manage.py runserver`

#### Showcase:
*Dashboard*
![Dashboard](https://raw.githubusercontent.com/realSaddy/Img2VecCosSim-Django-Pytorch/master/dashboard.png)

---

*Example Comparison*
![Example Comparison](https://github.com/realSaddy/Img2VecCosSim-Django-Pytorch/blob/master/example_comparison.png?raw=true)