Skip to content

lombokai/onepiece-classifier

Repository files navigation

Machine Learning Project Documentation

Project Overview

This repository contains programs for classifying images. The images used for this project are images of Onepiece anime characters. There are 18 predicted anime characters. This project is packaged in the form of a python package so that it can be used by the public.

Installation Guide

Requirements

It is recommended to run this application within a virtual environment. First thing first is clone the repository

git clone https://github.com/lombokai/onepiece-classifier.git

Virtual Environment Setup

create a virtual environment

python3 -m venv <virtual environment name>

command line to activate your virtual environment in linux

source <virtual environment name>/bin/activate

command line to activate your virtual environment in windows

<virtual environment name>\Scripts\activate

install package requirements

pip install -r requirements/main.txt

Usage

Example Usage

This package provide image to predict in assets directory. If you want to try predict an image, run predict.py script with command bellow

python3 predict.py <image path>

command example if you are working in parent directory of this repo

python3 predict.py assets/luffy.png

try to predict different image with command above

Run the Application

install onepiece-classify package

pip install onepiece-classify

acces predict method from onepiece-classify pakcage

from onepiece_classify.infer import ImageRecognition

instantiate the class with your model path. Download trained model here and make sure you specify model path parameter in the location of downloaded model

predictor = ImageRecognition(<model path>)

then you just predict your image with predict method

predictor.predict(<your image path>)

Data Description

Data Sources

Data obtained from Kaggle here. The data contains a collection of 18 onepiece character images, and is saved in jpg, jpeg and png formats.

Contributing

When contributing to this repository, please first discuss the change you wish to make via issue, email, or any other method with the owners of this repository before making a change.

License

License