The Car Connection Picture
Classification
Vehicle
|...
许可协议: Unknown

Overview

The Car Connection Picture Dataset is a dataset for car classificatioin.

Data Preview

Label Distribution

Instruction

HOW TO RUN

  1. Copy all .py files into a folder.
    • Make sure you have all the dependencies installed and up to date (e.g., bs4, requests, etc).
  2. In main.py set path to where the files are, and directory where you want the images to land
    • You do not need to create the directory yourself
  3. Runmain.py. I suggest you try it with a portion of the data first, in case an error emerges later.
    • For instance, in scrape.py line 27, replace for make in listed: to for make in listed[1:3]:

EXAMPLE. Example — Audi vs BMW ConvNet.ipynb : example of a deep learning classification task with Pytorch

WARNING: You may have issues if you use Python 3.6

FAQ

  1. How do I get the large pictures?
    • In scrape.py, row 68, change this line:
    • for ix, photo in enumerate(re.findall('sml.+?_s.jpg', fetch_pics_url)[:150], 1):
    • to this line:
    • for ix, photo in enumerate(re.findall('lrg.+?_l.jpg', fetch_pics_url)[:150], 1):
    • You can use sml, med, lrg for your preferred image size

FILES

FILES DESCRIPTION EXPORT
scrape.py Creates a df of all cars with their specs/pics URLs specs-and-pics.csv
tag.py Turns the previous df into one tag per URL id_and_pic_url.csv
save.py Turns all rows in the previous df to a picture named with the tag pictures/*.jpg
select.py Uses numpy to delete interior pictures, based on pixel color exterior/*.jpg
main.py Runs all other files None
数据概要
数据格式
Image,
数据量
64.467K
文件大小
681.38MB
发布方
Nicolas Gervais
Data Scientist at TD, former student in Machine Learning and Data Science at McGill, and in Business Analytics and Data Science at HEC Montréal.
数据集反馈
出错了
刚刚
timeout_error
立即开始构建AI
出错了
刚刚
timeout_error