README.md

---
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
dataset_info:
  features:
  - name: image
    dtype: image
  - name: label
    dtype:
      class_label:
        names:
          '0': octane
          '1': aftershock
          '2': werewolf
          '3': breakout
  splits:
  - name: train
    num_bytes: 6636053024.34
    num_examples: 380870
  download_size: 1429629384
  dataset_size: 6636053024.34
license: mit
task_categories:
- image-classification
tags:
- game
- rocket league
- mobile
- car
- synthetic
pretty_name: Rocket League Sideswipe
size_categories:
- 100K<n<1M
---

# Rocket League Sideswipe Vehicle Classification Dataset

![](https://git.fainsin.bzh/ENSEEIHT/projet-reseaux-profond/raw/branch/master/docs/demo_datas.gif)

## Dataset Description

- **Homepage**: https://git.fainsin.bzh/ENSEEIHT/projet-reseaux-profond
- **Report**: https://git.fainsin.bzh/ENSEEIHT/projet-reseaux-profond/src/branch/master/docs/rapport2.md

## Description

This dataset serves the purpose of vehicle recognition (classification) within the mobile video game Rocket League Sideswipe.
It comprises approximately 400,000 images.
The dataset was acquired through an automated script designed to customize in-game models (such as rims, hats, stickers, colors, ...)
and capture screenshots on an Android device, necessitating an approximate duration of 18 hours for compilation.